From preconception to postpartum: bidirectional associations between sleep and depression and the role of infant sleep in a population-based cohort

This study investigated bidirectional associations between maternal sleep and depressive symptoms from preconception to postpartum, and whether infant sleep mediated or moderated these associations. We used data from the Generation R Next Study (N=2,294). Maternal sleep (specifically general sleep disturbance, latency, quality, duration, and midpoint) and depressive symptoms were prospectively assessed at five timepoints from preconception to 12-month postpartum. Sleep was self-assessed with the General Sleep Disturbance Scale and Munich Chronotype Questionnaire; depressive symptoms with the Adult Self Report depression/anxiety subscale and Edinburgh Postnatal Depression Scale. Infant sleep (specifically night awakenings, nocturnal sleep duration, and latency) was parent-reported at 1-month postpartum using the Brief Infant Sleep Questionnaire. Bidirectional associations were examined using Autoregressive Latent Trajectory Models with Structured Residuals. The role of infant sleep was examined using mediation and moderation analyses. We found that maternal sleep and depressive symptoms were both stable over time. For sleep quality and disturbance, bidirectional associations suggested slightly stronger effects from depression to sleep (sleep quality:{beta}depression[->]sleep quality=0.11, 95%CI:0.07 - 0.14; general sleep disturbance:{beta}depression[->]sleep disturbance=0.14, 95%CI:0.10 - 0.18) than from sleep to depression ({beta}sleep quality/disturbance[->]depression=0.07 for both, 95%CIs:0.03 - 0.11). For latency, effects were comparable in both directions ({beta}depression[->]sleep latency=0.06, 95%CI:0.03 - 0.09; {beta}sleep latency[->]depression=0.05, 95%CI:0.01 - 0.09). The association between depressive symptoms and sleep latency was both mediated (9.7%) and moderated (p<0.05) by infant sleep latency. In conclusion, general maternal sleep disturbance, sleep quality, and sleep latency showed bidirectional associations with depressive symptoms from preconception/early pregnancy onwards. Infant sleep latency may represent a potential modifiable factor within this cycle.