Bibian van der Voorn

57 MATERNAL STRESS REFLECTED IN NEONATAL HAIR GC LEVELS were found between prenatal maternal stress scores and hair GC levels, we tested the relative contributions of prenatal and postnatal stress by analyzing combinations of (1) low pre- and low postnatal (reference), (2) low pre- and high postnatal, (3) high pre- and low postnatal or (4) high pre- and high postnatal stress. In addition, the influence of maternal SAD use was analyzed with hair GC levels as dependent factor, and SAD use as dichotomous (no medication vs. SAD use) or categorical independent variable (no medication use vs. use of SSRi, SNRI and/or NaSSA). Confounders were selected a priori, based on the literature 19-21,27 . Gender, birth weight (percentile) and primiparity were added to the multivariable model, one by one. Subsequently, based on statistical impact (>10% change in beta) the final model was created. When confounders were found to have a statistical impact on more than 50% of the associations being analyzed, we also explored their univariate influence on the outcome afterwards. RESULTS Characteristics of all subjects who donated hair pp are shown in Table 1 . Of these, a total of 66 women had consulted the POP center of expertise for pregnancy and psychiatric disorders, of whom 96% reported SAD use, 29% had an elevated HAS- subscore, and 17% had an elevated HDS-subscore. For mothers admitted for strictly medical reasons (n=103), these numbers (2%, 10%, and 8%, respectively) were similar to previously reported prevalence rates in normal populations 6-8,28 . Descriptive statistics for gender, gestational age, birth weight, primiparity, maternal age and ethnicity did not differ between the two subgroups.