Sleep Disturbance and Pregnancy Outcomes
Sleep disturbance, including short and long sleep duration and SDB, has been associated with increased mortality [ 55 ], hypertension [ 56 , 57 ], cardiovascular disease [ 58 ], obesity [ 59 ], altered glucose metabolism [ 60 ] and diabetes mellitus [ 61 ], and impaired immune cell activity [ 62 , 63 ] in nonpregnant populations. As early as 1990, a large questionnaire-based study described increased risk of preterm labor (11 % versus 6 %), preterm delivery (9.8 % versus 4.6 %), and preeclampsia (8.8 % versus 3.5 %) in a large registry of pregnant resident physicians compared to spouses of male colleagues [ 64 ]. Although sleep was not directly assessed, the difference in pregnancy outcomes was attributed to suspected sleep disturbance and deprivation, as the physician group worked signifi cantly longer hours than the control group (as high as 73 hrs/week in the first trimester in the physician group compared to 38 hrs/week in the non- physician group). Since then, a growing body of literature has addressed the relationship between sleep disturbance and pregnancy outcomes as discussed below.
Pregnancy Complications Gestational Hypertensive Spectrum
As has been described in nonpregnant populations, both short (<6 h) and long (>10 h) selfreported sleep durations were associated with 3.72 and 4.21 mmHg elevations in third trimester systolic blood pressures respectively, compared to systolic blood pressures in subjects with intermediate sleep durations (6-10 hrs/night), in a prospective cohort study of 1,272 pregnant women.
In a survey of 1,719 women in the third trimester of pregnancy, pregnancy-onset, but not chronic snoring was independently associated with gestational hypertension and preeclampsia but not gestational diabetes (GDM), although frequent snoring was shown to be associated with GDM in other studies.
In a retrospective cross-sectional study, Franklin et al. asked 502 women on the day of delivery whether they snored prior to pregnancy (4 %) or during the last week of pregnancy (23 % of women reported snoring every night during the last week). Gestational hypertension and preeclampsia developed in 14 % and 10 % of snorers, compared with 6 % and 4 % of non-snorers.
In a retrospective cohort study, Louis et al. found an association between PSG-confi rmed OSA in 57 women (median AHI 22, median oxygen saturation nadir 87 %, range 84–90 %) and incidence of preeclampsia (19 % in OSA patients versus 7 % in a control cohort of 115 women), as well as preterm birth (30 % versus 12).
However their study was complicated by preexisting comorbid characteristics of the OSA cohort, including higher BMI and chronic hypertension, as well as iatrogenically-induced medically-indicated preterm delivery in the context of preeclampsia.
Chen et al. also found increased risk of preeclampsia and preterm birth, but not GDM, in 791 pregnant women with confi rmed OSA. Finally, a recent study has found an independent association between OSA and preeclampsia even when confounding comorbidities such as BMI, maternal age, and diabetes were controlled for. Conversely, in a cross-sectional study, Reid et al. found higher prevalence of PSG-confi rmed OSA (defi ned as RDI >5) in pregnant women with gestational hypertension (53 %) compared to healthy pregnancy controls (12 %) [ 70 ].
Importantly, early OSA treatment with continuous positive airway pressure (CPAP) during the fi rst 8 weeks of pregnancy in women at risk for preeclampsia resulted in improved blood pressure control, lower antihypertensive medication requirements, and improved pregnancy outcomes.
In contrast to the fi ndings of O’Brien et al., frequent snoring (>3 nights/week) was independently associated with increased incidence of GDM in a prospective cohort study of 189 healthy nulliparous women, and so was self-reported short sleep duration (<7 h).
Qui et al. also describe an association between GDM and selfreported snoring and short sleep (<4 h). However their fi ndings also implicate long (>10 h) sleep duration, as well as elevated BMI (>25 kg/m 2 ) with increased incidence of GDM.
Lee and Gay used 48 hour actigraphy and sleep questionnaires in a prospective study of 131 women in the third trimester of pregnancy, and reported increased duration of labor and a 5.2 higher likelihood of cesarean delivery in women with actigraphy-derived WASO >15 % compared to WASO < 10 %, as well as longer labors and 4.5 times greater likelihood of cesarean delivery in women reporting <6 h of sleep compared to at least 7 h of sleep. Higher rates of cesarean delivery were also associated with self-reported poor sleep quality occurring more frequently than 3 days per week.
Higher cesarean rates have also been reported in association with PSG-confi rmed OSA during pregnancy, 65 % versus 4 % in the non-OSA pregnant controls, although higher BMI and incidence of chronic hypertension in the OSA compared to the control groups were signifi cant confounders.
In addition to an association between pregnancyonset snoring and gestational hypertension, Franklin et al. reported snoring as a risk factor for growth retardation of the fetus, with 7.1 % of small for gestational age (SGA) infants born to snoring mothers versus 2.6 % in non-snorers. APGAR scores were also <7 at 1 and 5 min, respectively, in 12 % and 3.5 % of babies born to snorers, compared with 3.6 % and 0.3 % of babies born to non-snorers. SGA and low Apgar scores at 5 min, as well as low gestational weight (LGW) have also been reported in babies of moms with confi rmed OSA. Finally, Louis et al. describe higher rates of neonatal intensive care unit (NICU) admissions of infants born to mothers with confi rmed OSA .
In summary, sleep disturbance has been described as a risk factor for adverse pregnancy outcomes, including increased incidence of gestational hypertension/diabetes, cesarean delivery, and longer labor in self-reported short sleep during pregnancy, as well as increased risk of gestational hypertension/diabetes, preeclampsia, higher cesarean rates, LGW, lower APGAR scores, and higher rates of NICU admissions in pregnancies complicated by SDB, especially in the OSA end of the SDB spectrum. Activation of infl ammatory pathways, including up- regulation of cytokine secretion, may represent the pathophysiological link between sleep disruption and adverse pregnancy outcomes. For example, Okun et al. report increased circulating and stimulated interleukin (IL)-6 levels in association with selfreported poor sleep quality (PSQI) and short sleep duration in mid- and late pregnancy, while IL-6 levels are elevated in preeclampsia.
Finally, despite the importance of sleep characterization in this population, there are no published studies defi nitively linking objective sleep measures (i.e. sleep stage as percent of TST, or arousal index) with maternal and neonatal outcomes. For example, although altered sleep architecture by PSG has been reported in preeclampsia , there are no prospective data on sleep disturbance as a cause rather than effect of preeclampsia. In conclusion, a growing research effort is underway to build upon and further refi ne the existing body of work describing sleep disruption during pregnancy and the complicated relationship between sleep parameters and pregnancy outcomes. As methodologies for objective sleep evaluation and proteogenomic characterization
become more sophisticated and widely available, longitudinal prospective studies with large numbers of participants will undoubtedly shed light on these important and fascinating topics.