Poster #8 - Predictors of change in breastmilk composition in the postpartum period: the case of cortisol

• In Event: Poster Session 13

Abstract

Background. Breastmilk is a primary source of nutrition with implications for infant’s cognitive and behavior development, and mother’s health (de Weerth et al., 2022). Next to nutritive factors (e.g. vitamins, carbohydrates), breastmilk contains nonnutritive bioactives (e.g. hormones, bacteria). Cortisol for instance, known as a stress hormone and the end product of the hypothalamic-pituitary-adrenal (HPA) axis, regulates metabolic, immunological and psychosocial challenges. Cortisol is absorbed in the infant gut (Pundir et al., 2017; Pacha, 2000). Pioneering work documented an increase in breastmilk cortisol concentrations (BCC) throughout the first three months postpartum (Browne et al., 2019). Other studies found an association between heightened BCC and neurobehavioral functioning (Hart et al., 2004), child body-mass-index (Hahn-Holbrook et al., 2016), and offspring fear behavior in humans (Glynn et al., 2021) and primates (Hinde et al., 2015), underscoring the potential implications of infant early exposure to alterations in BCC. While much research focused on the impact of breastmilk and mapped its complex composition, the factors contributing to (changes in) composition have been neglected. This is a fundamental question given that the postpartum period is knowingly challenging for new parents, due to biological, social, and relational changes (Law et al., 2018), with implications for mother’s biology and infant development. This study adopted a data-driven approach to identify maternal prenatal dispositional and postnatal situational factors that may explain BCC and their changes in the first three months postpartum. This study was preregistered.
Methods. Data from 73 mothers were collected prenatally, at birth, at 2-, 6- and 12-weeks postpartum. BCC was collected at 2-, 6- and 12-weeks postpartum. Predictors included: 1) dispositional factors, namely maternal own early experiences (i.e. attachment, empathy, adverse childhood experiences-ACEs); 2) situational factors, defined as aspects that emerge postpartum (i.e. psychological wellbeing, caregiving experiences and support, mother/father sensitivity, infant crying); and 3) additional factors known to influence cortisol (e.g. time of collection, season). The full list of tested variables is provided in Table 1. We performed two sets of regression models: 1) random forest to model interactions and non-linear relationships for each BCC assessment point, and 2) Bayesian linear models to capture predictors of changes in BCC over time (Figure 1).
Results. Random forest regression revealed that BCC was positively predicted by dispositional empathy and postnatal anxiety at week 2, and negatively predicted by ACEs at week 6, yet with poor explanatory power. Bayesian linear regressions with 50 imputed datasets also signalled that ACEs predicted BCC across 58% of datasets (Table 1). However, the effect size was small and the 95% credible interval overlapped with zero indicating that ACEs are a weak predictor of the change in BCC over time.
Discussion. Using sophisticated models, we found that no single factor had a strong predictive value of cortisol in breastmilk, beyond the known predictors of cortisol such as time since awakening and time of collection. Replication studies are recommended to give more weight to these findings. Noteworthy, both models suggest that ACEs are negatively associated with milk cortisol concentrations, warranting future research on the topic.

Authors

• Stefania V. Vacaru, Radboud University Medical Center
  Presenting Author

• Bonnie Erin Brett, Radboud Universiteit Nijmegen

• Henrik Eckermann

• Carolina de Weerth, Radboud University Medical Center