Later sleep timing is associated with an increased risk of preterm birth in nulliparous women.

backgroundAlthough uterine contractions have a diurnal periodicity and increase in frequency during hours of darkness, data on the relationship between sleep duration and sleep timing patterns and preterm birth are limited.

objectiveWe sought to examine the relationship of self-reported sleep duration and timing in pregnancy with preterm birth. STUDY

designIn the prospective Nulliparous Pregnancy Outcome Study: Monitoring Mothers-to-be cohort, women completed a survey of sleep patterns at 6-13 weeks gestation (visit 1) and again at 22-29 weeks gestation (visit 3). Additionally, at 16-21 weeks gestation (visit 2), a subgroup completed a weeklong actigraphy recording of their sleep. Weekly averages of self-reported sleep duration and sleep midpoint were calculated. A priori, sleep duration of <7 hours was defined as "short," and sleep midpoint after 5 am was defined as "late." The relationships among these sleep characteristics and all preterm birth and spontaneous preterm birth at <37 weeks gestation were examined in univariate analyses. Multivariable logistic regressions that controlled for age and body mass index alone (model 1) and with additional covariates (race, smoking, insurance, and employment schedule) following a backward elimination process (model 2) were performed.

resultsOf the 10,038 women who were enrolled, sleep survey data were available on 7524 women at visit 1 and 7668 women at visit 3. The rate of short sleep duration was 17.1% at visit 1 and 20.7% at visit 3. The proportion with a late sleep midpoint was 11.6% at visit 1 and 12.2% at visit 3. There was no significant relationship between self-reported short sleep and preterm birth across all visits. However, self-reported late sleep midpoint (>5 am) was associated with preterm birth . Women with a late sleep midpoint (>5 am) in early pregnancy had a preterm birth rate of 9.5%, compared with 6.9% for women with sleep midpoint ≤5 am (P=.005). Similarly, women with a late sleep midpoint had a higher rate of spontaneous preterm birth (6.2% vs 4.4%; P=.019). Comparable results were observed for women with a late sleep midpoint at visit 3 (all preterm birth 8.9% vs 6.6%; P=.009; spontaneous preterm birth 5.9% vs 4.3%; P=.023). All adjusted analyses on self-reported sleep midpoint (models 1 and 2) maintained statistical significance (P<.05), except for visit 1, model 2 for spontaneous preterm birth (P=.07). The visit 2 objective data from the smaller subgroup (n=782) demonstrated similar trends in preterm birth rates by sleep midpoint status.

conclusionSelf-reported late sleep midpoint in both early and late pregnancy, but not short sleep duration, is associated with an increased rate of preterm birth.