Background and objectives: Understanding vulnerability factors involved in the development of postnatal depression has important implications for theory and practice. In this prospective study, we investigated whether self-esteem instability during pregnancy would better predict postnatal depressive symptomatology than level of self-esteem. In addition, going beyond former studies, we tested the possible origin of this instability, examining whether day-to-day fluctuations in self-esteem could be explained by fluctuations in mood state, and whether this day-to-day self-esteem reactivity would predict postnatal depressive symptoms.

Methods: 114 healthy never-depressed women were tested during the late second or third trimester of their gestation (Time 1) and at 12 weeks after delivery (Time 2). Day-to-day levels of self-esteem and depressed mood state were assessed at Time 1. At Time 2, postnatal depressive symptoms were assessed.

Results: The results show that, after controlling for initial depressive symptomatology, age and socioeconomic status, postnatal depressive symptomatology at 12 weeks after childbirth could be predicted by self-esteem instability and not level of self-esteem. In addition, multi-level analyses demonstrated that these changes in day-to-day levels of self-esteem are associated with changes in day-to-day levels of depressed mood state and that those subjects with greater prenatal self-esteem reactivity upon depressed mood report higher levels of depressive symptoms post-partum.

Limitations: We used paper and pencil day-to-day measures of state self-esteem, which can be subject to bias.

Conclusion: These results provide evidence for a diathesis-stress account of postnatal depression, highlighting the importance of a multi-dimensional view of self-esteem and the predictive role of selfesteem instability.