ACTA ENDOCRINOLOGICA (BUC)

Context. Oxytocin (OXT) supports social affiliation and affect regulation; in late pregnancy, variability in plasma OXT may index the psychosocial context relevant to perinatal risk. Objectives. To test (i) differences in plasma OXT by adult attachment (AAS/R-AAS); (ii) associations with pregnancy planning, partner support, maternal perception, and history of emotional difficulties; and (iii) psychosocial factors independently associated with OXT. Design. Cross-sectional study. Subjects and Methods. Thirty-five women at 28– 39 weeks’ gestation (subsample of 140) provided morning EDTA plasma (08:00–10:00), stored at -80 ¯C. OXT was quantified by ELISA without solid-phase extraction or technical duplicates. Analyses comprised one-way ANOVA with Bonferroni tests, non-parametric contrasts (Mann– Whitney/Kruskal–Wallis with Holm adjustment), and HC3- robust multiple regression (α=0.05). Results. OXT differed by attachment (secure > avoidant > anxious–ambivalent; F(2,32)=31.115, p<0.001, ηp²=0.660). Univariate contrasts indicated higher OXT with planned pregnancy, partner support, positive (vs negative) maternal perception, and with a history of emotional difficulties; in the multivariable model (R²=0.718; adjusted R²=0.617), only attachment independently predicted OXT (avoidant vs secure B=-227.7 pg/mL; anxious–ambivalent vs secure B=-289.4 pg/mL). Conclusions. In late pregnancy, plasma OXT co-varies with adult attachment and, at group level, with proximal psychosocial factors; after adjustment, attachment remains the dominant predictor. Psychosocial screening is clinically actionable, whereas OXT measurement should remain a research tool.