Association between sleep duration during pregnancy and preterm birth: a dose-response meta-analysis

Objective Preterm birth is now a global health problem. There is a great alteration of sleep duration in pregnancy. Whether sleep duration in pregnancy affects preterm birth remains unclarified. This study aimed to examine the associations of sleep duration on preterm birth risk based on the dose-response meta-analysis. Materials and methods Electronic databases (PubMed, Web of Science, and Ovid) were searched for relevant studies from database inception to September 2020. Studies describing the associations of maternal sleep duration with preterm birth risk were included. A random-effects model was adopted to calculate the summarized relative risk (SRR) and 95% confidence intervals (CIs) as the effect sizes for all studies. Moreover, dose-response analysis was used for combining studies that used categories of sleep duration as exposure. In addition, subgroup analysis and meta-regression analysis were conducted to adjust potential confounders and investigate the source of heterogeneity. Results A total of 3771 unique studies were screened, and eight cohort studies and one case-control study were identified, with a total of 2000 preterm birth cases. Pooled data indicated that extreme sleep duration (short sleep duration or long sleep duration) in pregnancy was closely related to preterm birth in comparison with normal sleep duration (SRR = 1.13, 95%CI: 1.05-1.22) and there was no significant heterogeneity among studies (I-2 = 7.0%, p = .37). Begg's funnel plot failed to uncover any evidence of publication bias. The non-linearity in the association of sleep duration with preterm birth showed significance (p < .01). Considering pregnant women slept for 7 h as a reference, 4 h short sleepers had a higher predicted risk of preterm birth (RR = 1.10, 95%CI: 1.01-1.19). Additionally, as compared with women with normal sleep duration, the preterm birth risk with short sleep durations (SRR = 1.20, 95%CI: 1.05-1.37) was elevated among pregnant women, and long sleep duration was related to preterm birth after adjusting for age (SRR = 1.20, 95%CI: 1.01-1.42). Conclusion There is an association of extreme sleep duration in pregnancy with preterm birth. In a non-linear dose-response meta-analysis, a U-shaped relationship between sleep duration and risk of preterm birth was observed and pregnant women who slept 4 h/day had a significantly higher risk of preterm birth than those who slept normally.