The study looked at how stress during pregnancy changes tiny chemical tags on DNA (called methylation) in both mothers and babies, especially in genes that control stress hormones and the oxytocin system. The changes are small, vary a lot depending on the tissue and timing, and there’s no clear pattern that can be used in everyday health hacks yet.

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.