Stress-induced gene regulation in pregnant women: A systematic review and quantitative evidence synthesis

Prenatal psychosocial stress is linked to adverse maternal-infant outcomes, plausibly via stress-responsive endocrine pathways and epigenetic regulation. We examined associations between prenatal stress and maternal DNA methylation (DNAm) at key psychoneuroendocrine loci.

We conducted a PROSPERO-registered (CRD420251058768), PRISMA-aligned systematic review with quantitative evidence synthesis (QES). Searches (2014-2025) identified observational studies and one randomized trial sampling blood and saliva/buccal. Effects were standardized as Hedges' g and pooled with random-effects only for like-for-like subsets (k ≥ 2). Dependence was handled with multilevel models and robust variance estimation; risk of bias and certainty used ROBINS-I/RoB 2 and GRADE.

Findings for NR3C1 exon 1 F were heterogeneous across tissue, timing, and stress domain; outcomes were non-commensurable, so no single pooled estimate was computed. For FKBP5, intron 7 (k = 2) showed lower DNAm with greater trauma/PTSD burden (SMD = -0.36, 95 % CI -0.67 to -0.05; I² ≈ 0 %). Single-study contrasts suggested higher DNAm at intron 2 and lower DNAm at intron 5 with depressive symptoms. For OXTR (DNAm/mRNA), only single-study results were comparable; signals varied by phenotype/genotype. Absolute DNAm differences were typically < 1-3 %. Overall certainty was low to very low.

Prenatal stress is associated with detectable yet context-dependent epigenetic variation in peripheral tissues. No single CpG or region is ready for clinical use. Progress requires region- and tissue-matched replication, harmonized stress phenotyping with cell-type adjustment, and meta-analysis-friendly, open reporting.