Variability in perinatal sleep quality is associated with an atypical cortisol awakening response and increased mood symptoms.

Abstract

Objective: Pregnancy is often typified with a decrease in sleep quality, which for many women, progressively worsens across gestation and into the postpartum. A mechanism linking poor sleep with certain adverse pregnancy outcomes is dysregulation of the HPA axis resulting in atypically elevated cortisol production. While total cortisol output normally increases across pregnancy, the cortisol awakening response (CAR), a response to waking up, is influenced by factors such as stress and mood. It attenuates as pregnancy progresses, with normalization in the first weeks after delivery. The goals of the present study were to (1) assess the temporal relationship between sleep quality and cortisol indices across the perinatal period; (2) evaluate whether sleep quality was associated with postpartum mood; and (3) assess whether cortisol mediated these associations.

Method: Data were collected as part of the Healthy Babies Before Birth (HB3) study. Sleep quality, depressive and anxiety symptoms, and cortisol from four time-points (8-16 weeks gestation, 30-36 weeks gestation, 6 months postpartum, and 1-year postpartum) were assessed. Participants (N = 223) who had sleep quality (PSQI) and cortisol data from at least 1 of 4 time-points were included in analyses. Three salivary cortisol indices were calculated: cortisol awakening response (CAR), diurnal slope, and area under the curve (AUC). Multi-level models were run to predict cortisol parameters based on deviations and typical maternal sleep quality at each wave as well as mood outcomes.

Results: Multilevel (time, wave, and person) modeling indicated that sleep quality was not associated with any of the cortisol indices, and none significantly varied across time. However, when PSQI scores were higher than the woman's own mean sleep quality, the CAR slope was steeper (+1 point in PSQI, γ=0.18), and when PSQI scores were lower than mean, the CAR slope was flatter (-1 point, γ=0.11). Poorer sleep quality was associated with greater depression severity (γ = 0.367) and anxiety symptoms (γ = 0.120). Cortisol did not mediate the relationship between sleep quality and depression symptoms.

Discussion: Increases in PSQI scores, but not higher mean PSQI scores, were associated with a larger CAR. There was no association between sleep quality and the diurnal slope or AUC. These data suggest that variability in sleep quality is significantly associated with the amount of cortisol secreted upon awakening.