A mixed methods study of perinatal sleep and breastfeeding outcomes in women at risk for postpartum depression.

ABSTRACT

INTRODUCTION: Pregnant and postpartum women experience significant sleep disruption, but the role of perinatal sleep disturbances in breastfeeding is understudied. METHODS: In this observational cohort study, we used mixed methods to examine associations between perinatal sleep and breastfeeding. Forty-eight women (mean age 28.2 ± 4.9 years) who were euthymic at enrollment but had a history of major depression (n = 43) or bipolar disorder (n = 5) had sleep recorded with wrist actigraphy. We determined feeding status through daily diaries and used semi-structured interviews to identify themes regarding participants' experiences, breastfeeding decisions, and behaviors. To examine whether sleep disturbance during pregnancy predicted breastfeeding (BF) rates, we defined "lower sleep efficiency" (LSE) and "higher sleep efficiency" (HSE) groups based on the median split of actigraphic SE at 33 weeks' gestation (cutoff SE = 84.9%) and classified mothers as No-BF, Mixed-BF (BF + formula), and Exclusive-BF at 2 weeks postpartum. RESULTS: Percentages of women who did any breastfeeding were Week 2 = 72.3%, Week 6 = 62.5%, Week 16 = 50%. LSE mothers were less likely than HSE mothers to initiate breastfeeding (percent No-BF LSE = 45.8%, HSE = 16.7%, P < .05). Average actigraphic sleep onset, sleep offset, time in bed, sleep duration, and SE did not differ based on breastfeeding status at any time point. Qualitative themes included insufficient preparation for the demands of breastfeeding, interrupted and nonrestorative sleep, and unrelenting daytime tiredness. CONCLUSIONS: In our sample, preserved actigraphic SE during pregnancy was associated with initiation and continuation of breastfeeding. Future work should examine whether improving sleep in pregnancy improves mothers' feeding experiences.