Maternal sleep position has been associated with adverse pregnancy outcomes including stillbirth and fetal growth restriction.1–5 Sleeping on the back or right side is thought to increase the risk for adverse perinatal outcomes owing to compression of uterine blood vessels and decreased uterine blood flow.6 Because sleep position is a potentially modifiable risk factor, these observations have prompted public health campaigns.7 For example, the Tommy’s charity has a program termed the “Sleep On Side” campaign.7
Although prior data suggest an association between supine sleep and adverse pregnancy outcomes, the validity of these findings remains uncertain. Prior studies have included small numbers of women and importantly, interviews regarding maternal sleep position were conducted after the stillbirth or other adverse pregnancy outcomes occurred.1–5 This introduces the potential for considerable recall bias.
To address these prior methodologic limitations, we performed a secondary analysis of data from the Nulliparous Pregnancy Outcomes Study: Monitoring Mothers-to-Be (nuMom2b). This study included prospective data collection during pregnancy at two time points regarding maternal sleep position and provided detailed data regarding obstetric outcomes.
This was a secondary analysis of the nuMoM2b parent study and the Sleep-Disordered Breathing substudy, whose methods have been previously published.8,9 The study is registered at ClinicalTrials.gov under NCT01322529. Briefly, the nuMoM2b parent study was conducted at eight clinical sites and managed by an independent Data Coordinating and Analysis Center. Inclusion criteria for the parent study were nulliparity (no prior delivery as of at least 20 weeks of gestation) and a viable singleton pregnancy at the time of screening between 6 0/7 and 13 6/7 weeks of gestation. Enrollment occurred between October 2010 and May 2014.
The study was approved by the Institutional Review Boards at each clinical site and the Data Coordinating Center, and all participants gave written informed consent. The observational cohort study involved three study visits that were not part of clinical care. These occurred at 6 0/7–13 6/7 weeks of gestation (visit 1), 16 0/7–21 6/7 weeks of gestation (visit 2), and 22 0/7–29 6/7 weeks of gestation (visit 3). At each study visit, participants underwent clinical assessment, obstetric ultrasonogram, answered questionnaires, and had biospecimens collected. At visit 1 and visit 3 (but not visit 2), one of the questionnaires included questions about sleep position and other sleep parameters.8,9 These questions were similar to those used in prior pregnancy and sleep position studies and showed modest correlation with objective sleep position in one study.1,2,4,5,10 Specifically, women were asked what position they usually went to sleep in and woke up in during the last week and last night. Choices for sleep position included: on my left side mostly; on my right side mostly; both sides just as much; on my back mostly; on my front mostly; just as much on my side as on my front and back; and sitting up or propped up. A dichotomous sleep position variable was created by assigning a woman to the “left lateral sleep” category if she indicated that she slept “on my left side mostly”; otherwise she was assigned to “non–left lateral sleep.” Similarly, a dichotomous variable for supine sleep was created by assigning a woman to “supine sleep” if she indicated she slept “on my back mostly.” To group responses into four sleep position categories, if the woman selected “on my left side mostly,” she was assigned to “left lateral sleep”; if she selected “on my right side mostly,” she was assigned to “right lateral sleep”; if she selected “on my back mostly,” she was assigned to “supine sleep”; otherwise she was assigned to “other sleep position.”
The primary outcome for our analysis was a composite adverse pregnancy outcome including stillbirth, hypertensive disorders of pregnancy (mild, severe, or superimposed preeclampsia; eclampsia; or antepartum gestational hypertension), and a small-for-gestational-age (SGA) newborns. These were chosen because they all have been associated with abnormal placental function and most have been associated with non–left-sided sleep.1–5 Secondary outcomes were individual obstetric complications including stillbirth, hypertensive disorders of pregnancy, and SGA newborns. For all analyses, women with pregnancy losses before 20 0/7 weeks of gestation were excluded.
Precise and standard definitions of adverse pregnancy outcomes were used as previously described.8,9,11,12 Medical record abstraction was performed by trained research personnel using a standardized protocol. Also, detailed chart abstraction including assessment of blood pressure severity, new-onset neurologic disturbances, epigastric pain or pulmonary edema, and blood and urine laboratory results was done in all cases of hypertension or proteinuria by individuals blinded to sleep position.11 Women with suspected hypertensive disorders of pregnancy who presented atypically or were difficult to classify according to study criteria were adjudicated by the principal investigators, and final classification was reached by consensus.
Descriptive statistics were used to characterize the study population by adverse pregnancy outcomes status (yes vs no). Chi-square tests assessed adverse pregnancy outcome status with characteristics that were categorical, and analysis of variance F-tests were used for continuous measurements. Crude and adjusted odds ratios (ORs) and 95% Cis were calculated from univariate and multivariate logistic regression models to relate sleep position in early and mid-pregnancy (a time period that extends to 30 weeks of gestation, which includes the start of the third trimester) to adverse pregnancy outcomes. Adjustment covariates included maternal age (no older than 21, 22–35, and older than 35), body mass index (BMI [calculated as weight in kilograms divided by height in meters squared] under 25, 25 to under 30, 30 or greater), and chronic hypertension (yes, no) in early pregnancy. Midpregnancy analyses (occurring from 22 to 30 weeks of gestation) were also adjusted for rate of weight gain from early pregnancy to midpregnancy. As a sensitivity analysis, propensity score methods were also used to further account for confounding and to reduce bias in estimates of adjusted ORs (aORs).13
The analysis a priori defined “left lateral sleep” compared with “non–left lateral sleep” as the primary comparison of exposures because that is the sleep position recommended as most favorable for pregnancy. We also assessed supine sleep because this is the sleep position considered to be most unfavorable. Secondary analyses used different exposure categorizations: “left-sided sleep” compared with “supine sleep,” “left-sided sleep” compared with “right-sided sleep,” and combinations of sleep patterns.
An additional secondary analysis evaluated objective documentation of sleep position in a subset of participants who also underwent sleep-disordered breathing assessment, using a level 3 home sleep test.9,10 This sleep assessment took place immediately after visit 1 and visit 3, and used the Embletta-Gold device, which also recorded body position. For this analysis, a dichotomous sleep position variable was created based on whether the participant spent more than 50% of the time sleeping in the supine position.
All tests were performed at a nominal significance level of α=0.05, and all single degree of freedom tests were 2-sided. No correction was made for multiple comparisons. Analyses were conducted using SAS 9.3/9.4 software.
Of the 10,038 women enrolled in the parent study, this analysis included the 8,706 with data available for pregnancy outcomes as well as responses from at least one sleep questionnaire (Fig. 1). At baseline, women with unavailable data were slightly younger, more likely to be Hispanic or non-Hispanic black, more likely to have chronic hypertension, and less likely to have asthma than those with data available (Appendix 1, available online at http://links.lww.com/AOG/B515). Of the 3,705 women enrolled in the Sleep-Disordered Breathing substudy, there were objectively measured data available regarding sleep position for 3,133 at visit 1 and 2,474 at visit 3.
A total of 1,903 women had at least one adverse pregnancy outcome, with 178 having both SGA and hypertensive disorders, eight with stillbirth and SGA, three with stillbirth and hypertensive disorders, and two with all three complications (Table 1). Women with adverse pregnancy outcomes had higher BMIs, were more likely to smoke, more likely to have chronic hypertension, pregestational diabetes and kidney disease, were more likely to be non-Hispanic black, and were less likely to be non-Hispanic white than those without adverse pregnancy outcomes.
No self-reported sleep positions, including either non–left lateral or supine, were identified, either the night or the week before the study visit, that were associated with the composite adverse pregnancy outcome (Table 2).
Propensity score methods were used to adjust the results relating non–left lateral position in the week before the study visit to the composite outcome for all variables in Table 1 (results not shown). Again, no association was found.
As with position at sleep onset, there was no association of self-reported sleep position on awakening (including non–left lateral or supine sleep) the night or the week before the study visit with adverse pregnancy outcomes (Table 3).
There was no association between any self-reported sleep position, including non–left lateral and supine sleep position, and gestational hypertensive disorders or SGA newborns (Table 4). Results were unchanged when SGA newborn was defined as having a weight of less than 5% for gestational age (data not shown). Propensity score methods were also used to adjust results relating non–left lateral position in the week before the study visit to individual adverse pregnancy outcomes of hypertensive disorders and SGA newborns for all variables in Table 1 (results not shown). Again, no association was found. Similar results were obtained regarding self-reported sleep position on waking up and individual adverse outcomes (data not shown).
Non–left lateral sleep in midpregnancy was associated with a decreased risk of stillbirth (5/4,667 [0.1%] vs 13/3,511 [0.4%]; aOR 0.27; 95% CI 0.09–0.75). Results of sensitivity analyses for the stillbirth outcome using propensity score matching were similar (aOR 0.28; 95% CI 0.09–0.90). We did not analyze further subsets of sleep position and stillbirth owing to the small number of stillbirths (n=18) that could be included in the analysis at visit 3. Although there were 40 women with stillbirth who completed a visit-1 or visit-3 sleep questionnaire, only 10 occurred at at least 37 weeks of gestation and 18 occurred at no more than 24 weeks of gestation.
The relationship between objectively measured sleep position and adverse outcomes was similar to that for self-reported sleep position (Table 5). Women with supine sleep position for more than 50% of the time were no more likely than those in the supine position no more than 50% of the time to have the composite adverse outcome. When other thresholds were explored, there was no relationship between increasing amounts of time spent in supine sleep and adverse outcomes (Appendix 2, available online at http://links.lww.com/AOG/B515). There were too few individuals with each individual outcome in this subset to allow for meaningful analysis of objective sleep position and individual adverse outcomes.
We also explored the relationship between self-report and objectively measured sleep position. There was modest correlation between subjectively reported and objectively assessed sleep position. For example, women who reported sleeping on their backs were more likely to spend more than 50% of time in objectively measured supine sleep than those who did not (60.7% vs 30.3% visit 1; 51.6% vs 48.4% visit 3).
Post hoc, we considered the ORs that could be detected with the sample sizes and frequencies of adverse pregnancy outcomes we had observed. Using a sample size of 8,000 participants, a significance level=0.05; power=0.80; a probability of non–left lateral sleep=.57; and the probabilities of adverse pregnancy outcomes given left lateral sleep equal to 13% for hypertensive disorders of pregnancy, 10% for SGA, 0.4% for stillbirth, and 22% for the composite. We found that we could detect ORs of 1.2 for hypertensive disorders of pregnancy, 1.23 for SGA, 2.4 for stillbirth, and 1.2 for the composite. That is, our study was adequately powered to detect clinically meaningful ORs.
Going to sleep in the supine or right lateral position was not associated with an increased risk of a composite outcome such as stillbirth, SGA newborn, and gestational hypertensive disorders compared with going to sleep in the left lateral position. This was true for sleep position during the last night or prior week in both early and midpregnancy. A null finding was also shown using objective data on sleep position in a subset of the cohort.
These results differ substantially from several prior studies. In a case-control study in New Zealand that included 155 women with stillbirths and 310 other women in the study, women who slept on their backs or right sides had an increased risk of late stillbirth (aOR 2.54 [95% CI 1.04–6.18]) compared with those who slept on their left sides.1 The absolute risk of late stillbirth for left side sleepers was 1.96 of 1,000 compared with 3.93 of 1,000 for non–left-side sleepers.
Similar results were noted in a case-control study in Australia. Supine sleeping was associated with an aOR of 6.26 (95% CI 1.2–34) for late stillbirth in 103 women with stillbirths and 192 other women in the study.2 A cross-sectional study of 232 women in Ghana noted an increased risk of stillbirth (OR 8.0; 95% CI 1.5–43.2) and low birth weight (OR 5.0; 95% CI 1.2–20.2) in women with supine sleeping compared with other sleep positions.3 Sleep position was self-reported after delivery.3 Another case-control study of 164 women with late stillbirth noted an aOR of 3.67 (95% CI 1.74–7.78) for supine going to sleep position the night before stillbirth.5 Finally, a recent multicenter case-control study in the United Kingdom that included 291 third-trimester stillbirths noted an aOR of 2.3 (95% CI 1.04–5.11) in women going to sleep in the supine position.4 The latter two studies used gestational age–matched ongoing pregnancies as controls.
The major difference between these studies and the current one is the timing of ascertainment of sleep position. Each of these studies asked women about sleep position after the stillbirth or other adverse outcome occurred, leaving the potential for bias. This is especially true in cases of stillbirth, because many women may have heard or been told to “not sleep on their backs” during pregnancy. Although more recent studies attempted to reduce recall bias by using structured questionnaires with questions about many different factors, the potential for recall bias, perhaps linked to parental guilt is considerable. In contrast, our prospective cohort design allowed for ascertainment of sleep position before the occurrence of the pregnancy outcome. Other differences among investigations included populations studied (nulliparous vs multiparous), single compared with serial assessment of sleep position, accuracy of gestational dating criteria, and timing in gestation of sleep assessment (late first and early third trimesters vs mid-late third trimester). Finally, we noted similar results using objective measures of sleep position in a subset of women.
It is proposed that supine sleep may compress the vena cava and aorta, potentially reducing venous return and ultimately, placental blood flow. Decreased maternal cardiac output, maternal hypotension, and reduced fetal oxygenation have been reported with supine and right-sided maternal position compared with left.14–16 In addition, a recent study noted an increase in abnormal fetal behavioral states as assessed by fetal heart rate tracings in maternal supine or right lateral compared with left lateral position.17 Thus, it is biologically plausible that supine or right-sided sleep could increase the risk of stillbirth or other adverse outcomes associated with decreased placental blood flow. Another factor is an increase in the risk for sleep-disordered breathing in the supine position.18 Sleep-disordered breathing also is associated with adverse pregnancy outcomes such as preeclampsia.11
Maternal sleep position is an attractive pathway for stillbirth and placental insufficiency, because it is a potentially modifiable risk factor for stillbirth and other adverse pregnancy outcomes. Accordingly, public health efforts and educational programs could lower the rate of adverse pregnancy outcomes in a manner similar to safe sleeping promotion and sudden infant death syndrome. Nonetheless, such efforts have a potential downside. Some women may have trouble sleeping on their left side and they cannot control movement during sleep. Even with careful messaging, there is potential to increase anxiety in women who wake up on their backs and guilt, shame and self-blame in women suffering adverse pregnancy outcomes such as stillbirth. There is even the potential for cost and harm if anxiety leads to unnecessary antenatal testing and false positive results. Thus, it is critical to be sure that the benefits outweigh risks before implementing widespread public health campaigns.
A limitation of our study was the relatively small number of stillbirths, especially in late pregnancy, which limited our ability to adequately assess confounding. Accordingly, we cannot definitively exclude an association between supine sleep and late stillbirth owing to limited sample size. However, the direction of the signal was in the opposite direction, with non–left lateral sleepers actually having a statistically lower risk of stillbirth compared with left lateral sleepers. (It is unlikely that left-sided sleep increases the risk for stillbirth, and this observation is likely spurious owing to small numbers.) Also, we were unable to objectively document sleep position in all participants. Others have noted good but imperfect correlation between recalled and actual going to sleep position.10,19,20 Two small studies of videotaped sleep reported correlations for sleep onset position and patient recall of 0.48–0.52.10,19 An observational cohort study comparing self-recall and polysomnography reported sleep position found significant correlation between perceived and objectively measured supine sleep (r-square=0.63, P<.001).20 However, there was poor negative predictive value for self-report and supine sleep and estimates of supine sleep duration were inaccurate.20 We also noted relatively modest correlation between self-reported and objective sleep position. Thus, future studies should aim to objectively assess sleep position and duration when possible. Our objective measurements also did not distinguish between left and right lateral positions.
Importantly, we did not assess sleep position in the last 2 months of pregnancy. This may be the most sensitive window for supine sleeping to adversely affect pregnancy. Accordingly, our data do not exclude a relationship between non–left lateral or supine sleep at the end of pregnancy and late stillbirth.
Our study also had numerous strengths. Prospective assessment of sleep position before delivery considerably reduces recall bias. In addition, our study has a relatively large sample size with considerable racial, ethnic, and geographic diversity. We also had excellent gestational dating criteria, granular data regarding participants, and rigorous ascertainment of pregnancy outcomes and assessment of sleep habits at two time points. Finally, we had objective assessment of sleep position in a large subset of pregnancies. Given the relative infrequency of stillbirth and considerable expense, it will be difficult to perform larger prospective studies of the relationship between sleep position and this adverse pregnancy outcome.
In summary, supine or non–left-sided sleep through 30 weeks of gestation was not associated with adverse pregnancy outcomes linked to decreased placental blood flow in a large prospective cohort. These data should provide reassurance to women regarding sleep position through 30 weeks of gestation.
1. Stacey T, Thompson JMD, Mitchell EA, Ekeroma AJ, Zucollo JM, McCowan LME. Association between maternal sleep practices and risk of late stillbirth: a case-control study. BMJ 2011;34:d3403.
2. Gordon A, Raynes-Greenow C, Bond D, Morris J, Rawlinson W, Jeffery H. Sleep position, fetal growth restriction, and late-pregnancy stillbirth. Obstet Gynecol 2015;125:347–55.
3. Owusu JT, Anderson FJ, Coleman J, Oppong S, Seffah JD, Aikens A, et al. Association of maternal sleep practices with pre-eclampsia, low birth weight, and stillbirth among Ghanian women. Int J Gynaecol Obstet 2013;121:261–5.
4. Heazell A, Li M, Budd J, Thompson J, Stacey T, Cronin RS, et al. Association between maternal sleep practices and late stillbirth—findings form a stillbirth case-control study. BJOG 2018;125:254–62.
5. McCowan LME, Thompson JMD, Cronin RS, Gordon A, Raynes-Greenow C, Heazell AEP, et al. Going to sleep in the supine position is a modifiable risk factor for late pregnancy stillbirth: findings from the New Zealand multicenter stillbirth case-control study. PLoS One 2017;12:e0179396.
6. Ryo E, Okai T, Kozuma S, Kobayashi K, Kikuchi A, Taketani Y. Influence of compression of the inferior vena cava in the late second trimester on uterine and umbilical artery blood flow. Int J Gynaecol Obstet 1996;55:213–8.
8. Haas DM, Parker CB, Wing DA, Parry S, Grobman WA, Mercer BM, et al. A description of the methods of the nulliparous pregnancy outcomes study: monitoring mothers to be (nuMoM2b). Am J Obstet Gynecol 2015;212:539.e1–24.
9. Facco FL, Parker CB, Reddy UM, Silver RM, Louis JM, Basner RC, et al. NuMoM2b Sleep-Disordered Breathing study: objectives and methods. Am J Obstet Gynecol 2015;212:542.e1–127.
10. Warland J, Dorrian J. Accuracy of self-reported sleep position in pregnancy. PLoS One 2014;9 e115760.
11. Facco FL, Parker CB, Reddy UM, Silver RM, Koch MA, Louis JM, et al. Association between sleep-disordered breathing and hypertensive disorders of pregnancy and gestational diabetes. Obstet Gynecol 2017;129:31–41.
12. Esplin MS, Elovitz MA, Iams JD, Parker CB, Wapner RJ, Grobman WA, et al. Predictive accuracy of serial transvaginal cervical lengths and quantitative vaginal fetal fibronectin levels for spontaneous preterm birth among nulliparous women. JAMA 2017;317:1047–56.
13. D'Agostino RB Jr. Propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group. Stat Med 1998;17:2265–81.
14. Kauppila A, Koskinen M, Puolakka J, Tuimala R, Kuikka J. Decreased intervillous and unchanged myometrial blood flow in supine recumbency. Obstet Gynecol 1980;55:203–5.
15. Milsom I, Forssman L. Factors influencing aortocaval compression in late pregnancy. Am J Obstet Gynecol 1984;148:764–71.
16. Cluver C, Novikova N, Hofmeyr GJ, Hall DR. Maternal position during cesarean section for preventing maternal and neonatal complications. The Cochrane Database of Systematic Reviews 2010, Issue 6. Art. No.: CD007623. DOI: 10.1002/14651858.CD007623.pub3.
17. Stone PR, Burgess W, McIntyre JP, Gunn AJ, Lear CA, Bennet L, et al. Effect of maternal position on fetal behavioural state and heart rate variability in healthy late gestation pregnancy. J Physiol 2017;594:1213–21.
18. Barnes H, Edwards BA, Joosten SA, Naughton MT, Hamilton GS, Dabscheck E. Positional modification techniques for supine obstructive sleep apnea: a systematic review and meta-analysis. Sleep Med Rev 2017;36:107–15.
19. McIntyre JP, Ingham CM, Hutchinson BL, Thompson JM, McCowan LM, Stone PR, et al. A description of sleep behavior in healthy late pregnancy, and the accuracy of self-reports. BNC Pregnancy Childbirth 2016;16:115.
20. Wallbridge PD, Churchward TJ, Worsnop CJ. Accuracy of patient perception of supine sleep. J Clin Sleep Med 2018;14:1205–8.