Abstract
Background
Sleep-disordered breathing (SDB), is an umbrella term that encompasses obstructive sleep apnea (OSA), central sleep apnea (CSA) and hypoventilation. is common but studies in the pregnant population are limited. Data suggests relationships between OSA and preeclampsia, but the relationship between snoring and pregnancy outcomes is unknown.
Methods
A prospective study of 2224 singleton pregnancies was undertaken. Women were questioned using the Berlin Questionnaire (BQ- 2 or more categories where the score is positive.) and the Epworth Sleepiness Scale (ESS >10/24), the results compared with pregnancy outcomes with regard to hypertension in pregnancy.
Results
Women having symptoms raising the possibility of OSA defined by the BQ with a score >7 was 45.5%, and using ESS with a score >10, was 36%. The birth and neonatal outcomes for self-reported snoring and increased daytime sleepiness showed increased adverse outcomes notably increased caesarean section rates and low APGAR scores but not birth before 37 weeks of gestation.
Conclusion
Using questionnaires designed for the general population, the prevalence of possible undiagnosed OSA is high in the pregnant population. The increased adverse delivery and neonatal outcomes for self-reported snoring and increased daytime sleepiness with these tools indicated the need for further investigation of the links between snoring SDB and pregnancy outcomes.
Introduction
Obstructive sleep apnea (OSA) consists of a spectrum of symptoms and signs ranging from snoring, to recurrent episodes of upper airway collapse and obstruction during sleep associated with recurrent oxygen desaturations and arousals from sleep. 1 It is well known that OSA is associated with daytime somnolence (sleepiness), trouble sleeping at night, snoring, and hypertension 1 . The modified Berlin questionnaire (BQ) is a standardised validated self-reported patient questionnaire designed to identify snoring and patients with sleep apnea. 2 The Epworth Sleepiness Scale (ESS) is a self-administered patient questionnaire with 8 questions to provide a measure of a person's general level of daytime sleepiness or average sleep propensity in daily life. 3 It is a well-accepted tool in the assessment of patients suspected of having OSA and its consequences 3 in the non-pregnant population, but the value of these cheap and accessible tools in pregnancy is unclear.
Changes that occur during pregnancy such as weight gain and diaphragm displacement can enhance the likelihood of the mother developing OSA. 4 Other factors further increase the risk of OSA such as hormone changes which induce hyperaemia, nasopharyngeal mucosal oedema, 5 and vasomotor rhinitis, 6 which can lead to a narrowing of the upper airway. The resultant increased resistance to airflow has been shown to be an independent risk factor for OSA in the general adult population7,8 and is highly likely to be relevant in pregnancy.
In pregnancy, self-reported snoring has been variably reported to occur in between 14 and 45% of pregnant women.9,10 There is however considerable controversy about whether snoring is associated with adverse maternal or fetal outcomes,10–12 and whether the current tools have relevance in pregnancy. Snoring questions have only been tested in their predictability for OSA and not for pregnancy outcomes. 13
Excessive sleepiness with progression to insomnia in pregnancy has been termed by the American Sleep Disorders Association as “pregnancy-associated sleep disorder”. 14 The frequency of this disorder is reported between 66%-94%.15,16 First trimester women report daytime sleepiness, insomnia and repeated nocturnal waking, resulting in a decrease in sleep efficiency. 15 Although these symptoms may continue throughout the pregnancy, 20% resume pre-pregnant sleep patterns in the second trimester. 16 During the third trimester however, increased insomnia and increased frequency in nocturnal waking was noted as well as a decrease in daytime alertness. 16 There is a lack of good quality correlations between the symptoms that are reported and whether any of the symptoms are associated with adverse birth outcomes.
The present study was undertaken to test whether scores from the modified BQ for risk categorization of snoring or from the ESS to assess sleepiness scores were associated with poor delivery and neonatal outcomes.
Methods
Women attending all antenatal clinics at Campbelltown Hospital, NSW, Australia, were eligible to participate in the study. Campbelltown is an outer metropolitan hospital in Sydney Australia which supports 3900 pregnancies per year. The study was approved by the Institutional Human Ethics Committee (South Western Sydney Local Health District) and women enrolled in the study gave written and informed consent. All pregnant women presenting for antenatal care were offered to participate in the study regardless of gestational age of the current pregnancy. Women were excluded from the study for multiple pregnancies only. They were asked to complete the ESS and BQ questionnaires regarding their sleep, snoring and daytime sleepiness while pregnant. The women were defined as “self-reported snorers” if they answered positively to the BQ question “Do you snore?” (Category 1 question). Category two questions (BQ) consisted of those related to tiredness and fatigue, and the Category three question regarded body mass index (BMI) (the modified questionnaire). The ESS rated on a 4 point (0–3) scale their usual chance of falling to sleep in 8 different situations of daily life. The total ESS score was the sum of the 8 items and can range from 0–24. The higher the score, the higher the daytime sleepiness level 4 and total scores were used in this study. Characteristics such as BMI (kg/m2), neck circumference and blood pressure were determined by standard measurements. Gestation diabetes mellitus (GDM) was defined as a positive result in the 75g oral glucose tolerance test (OGTT) during the pregnancy.
Patients completed the forms however the staff were available to assist if required, usually to assess in clarification of the questions. Clinical characteristic data was obtained from the patient's electronic medical records. Anthropometric measures were undertaken at the initial visit and included neck circumference (measured at the level of the cricoid cartilage and rounded to the nearest millimetre), as were blood pressure (BP), gestational age and pregnancy details. First, second, and third trimesters were defined as first trimester under 12 weeks and 6 days of gestation; second trimester 13 weeks to 28 weeks and 6 days of gestation; and third trimester of 29 weeks or more of gestation.
Pregnancy maternal and neonatal outcomes birth weight (g), length and head (cm), APGAR scores at 5 min and delivery method were obtained from the hospital database and medical records at the time of delivery. Where possible, the date for those who delivered at other institutions was sought with participant permission or, via participant recall (<2%). The diagnosis of preeclampsia and gestational hypertension was based on the Society of Obstetric Medicine of Australia and New Zealand criteria. 17 This required blood pressure of over 140mmHg systolic or 90 mmHg diastolic and for preeclampsia, additional critical features including proteinuria. Outcomes included gestational at delivery, mode of delivery and transfer of the newborn to the neonatal intensive care unit (NICU).
Statistical Methods
Demographical and clinical data were compared by student T-test Categorical data were compared by chi-square and p < 0.05 was used for statistical significance. The modified BQ and ESS were correlated by linear regression. The different categories of BQ questions were analysed for risk of hypertension in pregnancy (all causes) (reported as odds ratio). The data are presented as means ± standard deviation (SD) for continuous variables and as rates for categorical values. Where relevant, non-normally distributed outcomes were compared by Mann Whitney-U test and presented as median ± inter quartile range. Multiple regression modelling controlled for maternal age, weight, and BP entry.
Results
A total of 2309 Berlin and Epworth questionnaires were completed. Women who had multiple gestations (43) or were lost to follow-up (26) or had a detal loss under 20 weeks (16) were excluded leaving 2224 questionnaires (Figure 1).
Consort diagram of participant flow from questionnaire to birth with less than 2% lost to follow-up.
The baseline characteristics of the women enrolled are detailed in Table 1. The median age of participants was 28 (24–32) years of age and median gestation at completion of the questionnaire was 28 (21–35) weeks. The pre-pregnant BMI was elevated to obese 29.9%, however at the time of completion of the questionnaire, 45.2% were obese (BMI ≥ 30.0 is obese). There were 461 (20.4%) women who smoked cigarettes. Gestational diabetes was diagnosed (with a positive result 75g OGTT) in 295 women (13.0%), hypertensive disorders complicated 248 (11%) pregnancies. Specifically, gestational hypertension occurred in 41 (1.8%) pregnancies, preeclampsia in 118 (5.2%), chronic hypertension in 56 (2.5%) pregnancies and 33 (1.5%) pregnancies were complicated by superimposed preeclampsia.
Descriptive clinical characteristics for women included in questionnaire analysis following birth.
IQR: interquartile ratio; SBP: systolic blood pressure; DBP: diastolic blood pressure; BMI: body mass index; IUGR: intrauterine growth restriction.
The snorers and non-snorers are listed; the women which answered “don’t know” to snoring are not in the table.
Category one (Do you snore?) of the Berlin questionnaire was positive in 41% (836) of the participants which is 2.5 times higher than that observed in women in the US study, despite the fact that women in our study may have been younger. Self-reported snoring was noted by 45.8% of participants. Of all snorers, 25.7% admitted to snoring very loudly and 38.8% stated their snoring was as aloud as breathing. The snoring of 699 women was known to have bothered other people. Women (n = 17) reported that others had noted their breathing stopped almost daily, breathing stopped 3–4 times a week in 12 women and 1–2 times a week in 33 women. Loud snoring is a hallmark of OSA and is known to affect up to 25% of all women in the United States, according to the U.S. National Institutes of Health. 18
Category two (related to tiredness and fatigue) questions of the Berlin questionnaire were positive in 48% of respondents. Fatigue or feeling tired after sleep was noted in 41% of respondents. Although 54.9% had noted fatigue whilst awake, very concerningly 96 women reported falling asleep whilst stopped at lights or in traffic when behind the wheel of a vechile.
High risk of OSA (2 or more categories positive in the BQ) occurred in 34.1% of women and Table 2 demonstrates the proportion of women with positive responses for each section of the Berlin questionnaire. This was associated with an increased risk of developing a hypertensive disease of pregnancy (HDP): OR 1.95 (95% CI: 1.45–2.63). There was a significantly increased risk of developing a hypertensive disorder of pregnancy if women had a positive result for Category 1 or Category 3 questions. (Table 2)
Berlin questionnaire results.
Positive category 1 = Do you snore? Yes, positive category 2 = a score of > 9 out of 24 in the Berlin questionnaire, positive category 1 & 2 is yes to question 1 and > 9 as a total score for the Berlin questionnaire. Category 3 was positive if the pre pregnant BMI Kg/m2 was > 30. High risk for SDB if pre pregnant BMI Kg/m2 was > 30 and increased risk of SDB if current pregnancy BMI Kg/m2 was > 30. BMI: body mass index); SDB: sleep disordered breathing.
The self-reported snorers were older by 2 years, had higher diastolic blood pressure (DBP) and systolic blood pressure (SBP) and higher BMI (Table 1). The women were enrolled at an average gestational age of 28 ± 5 weeks, were 28 ± 5 years old. Of the cohort 45.5% were self-reported snorers. Women that were self-reported had significantly higher BMI (27 vs. 24.6, p < 0.0001) and higher booking-in blood pressures when the women are categorised into self-reported snorers and self-reported non snorers.
Those who self-reported snoring (yes at questionnaire to question 1) had a total Berlin questionnaire (BQ) score mean (SD) of 7 (±2) and non-snorers (no at questionnaire to question 1) had a mean (SD) of 5 (±1.7), which was highly significantly different (p < 0.0001) (Table 1). Of note, 27.5% scored >5 and would normally be considered for further investigation and possible treatment with CPAP (Figure 2). The analysis of frequency of overnight snoring questions indicated that 39% of all pregnant women snore in excess of 3 nights per week.
This demonstrates the spread of total Berlin scores in this pregnant population. The red arrow shows the treatment cut off score as required by the general population. There were 27.5% of this pregnant cohort who scored in the usual treatment range (>5).
The mean ESS (SD) was 7(±4.1) with a range of 0–24. Those who self-reported snoring had a total ESS mean (SD) score of 8 (±4.2) and non-snorers a total ESS of 6 (±4), (p < 0.0001) (Figure 3).
This chart demonstrates the propensity to fall asleep scores for the cohort with 35.8% meeting intervention criteria based in the general population. In the non-pregnant population the score cut off for mild and severe treatment requirements are represented by the red arrows.
Of the eight component situations of ESS, the individual item propensity for “falling asleep while resting in the afternoon” showed, 77.9% of participants who nap 3 or more times a week, 34.7% while “sitting and reading” and 53.4% while “watching television”.
Among the responses, were very concerning number (96) of pregnancies “who fall asleep while at the wheel of a motor vehicle (stopped at lights or in traffic)”. Compared to one in 25 American adults who fall asleep while driving. There were 3.7% of 2266 women who admitted to falling asleep while driving (stopped at lights or in traffic) and of that, 7.1% stated they fall asleep every day while behind the wheel stopped in traffic or at lights. In all 35.8% would qualify for further testing for OSA and intervention based on a score >9 (Figure 3).
There was a decrease in vaginal births for those women who self-report snoring (61%) versus non-snoring (65%) due to an increase in the caesarean section rate (31% of self-reported snorers and 27% non-snorers (p < 0.001)) (Figure 4). Snorers had a significantly greater chance of low APGAR score at birth but no difference in rate of delivery prior to 37 weeks of gestation.
Birth outcomes comparison between snorers and non-snorers. NVD = normal vaginal delivery, C/S = caesarean section, Assist VD = vacuum or forceps vaginal delivery, SCN = special care nursery (multiple comparison and pair wise comparison). Multiple regression modelling controlled for maternal age, weight and BP entry.
Discussion
Obstructive sleep apnea (OSA) is a common disorder characterized by repetitive narrowing or collapse of the pharyngeal airway during sleep causing snoring and apnea events. The disorder is associated with major comorbidities including excessive daytime sleepiness and increased risk of cardiovascular disease in the general population. In the pregnant population OSA is a modifiable risk factor for adverse maternal and infant conditions. The underlying pathophysiology is multifactorial and may vary considerably between individuals and gestations.
These are the findings from a large prospective study to look at the pregnancy outcomes in pregnant women who self-report snoring using the Berlin questionnaire the ESS. The outcomes from this study demonstrate that our women have a higher prepregnant BMI, an increased risk of hypertensive disease in pregnancy, snorers are older and deliver earlier with increase infants with <7 Apgar scores and higher caesarean section rates. Louis and colleagues found that age, BMI, and frequent snoring predicted OSA, however their study was of nullipara low risk women only. Facco and colleagues 19 have tested models in high risk women, with self-reported frequent snoring, chronic hypertension, BMI, and age emerging in one study, whereas Wilson's group tested for snoring alone.
The use of self-reported snoring as a tool to detect OSA is well established. In this large prospective study, symptoms of OSA were common in our randomly selected prospective evaluation of pregnant women. Regardless of trimester, the presence of snoring (positive category one questions “Do you snore?”) was strongly correlated with the risk of hypertensive disease of pregnancy. A recently published large cohort study found that new-onset snoring during pregnancy is a strong risk factor for gestational hypertension and preeclampsia. 14
This study demonstrated OSA rates in this pregnant population are in excess of 45% when using the BQ and 38% with the ESS, which is above those demonstrated by Schultze (tool; score), Pien et al. 20 tool; score and Edwards tool; score who only utilised one of the questionnaires. This increase in OSA prevalence can be attributed to the increased BMI within this community but foreshadows what's to come in the future with the increased weight of the population as a whole. The prevalence of self-reported snoring was associated directly with greater gestational age and BMI, both pre-pregnant and recruitment pregnancy weight. The <7 Apgar's was associated with poorer outcomes. There was a decrease in vaginal births for those women who self-report snoring, the caesarean sections were increased in those who self-reported snoring. The Berlin questionnaire is self-reported; snoring is a component of the Berlin questionnaire (Adapted from: Table 2 from Netzer, et al., 1999) which has been validated in men and post-menopausal women but not pregnancy.21–23
Within this pregnant population, a total 35.8% of participants would in normal circumstances seek medical advice for sleep apnea. Those requiring further testing and treatment intervention such as CPAP are represented in the severe OSA accounts for 8.8% of this pregnant cohort. The mild to moderate cut off is also demonstrated whereby 27% of this cohort could seek medical advice regarding OSA. The category of snoring performed in predicting poor outcome as evidence by increased rates of poor initial newborn well-being scores and rates of caesarean delivery.
The risk factors for OSA are well established in the general population and include: male gender, older age, obesity, African-American race, craniofacial abnormalities, and smoking.24,25 OSA is also associated with other co-morbid conditions including type II diabetes, hypertension, cardiac arrhythmias, and cardiovascular disease. 26 Given that our women were heavier, older and had higher BP, the physiological changes of normal pregnancy predisposing to OSA include upper airway oedema and respiratory and hormonal driven changes may be enhanced. Many women may enter pregnancy with snoring and develop deteriorating airway obstruction due to physiological and hormonal changes of pregnancy or in association with other comorbidities developed in pregnancy (hypertensive disorders of pregnancy, or gestational diabetes).
The effects of OSA on pregnancy are an emerging field. The fetal outcomes from mothers who snore during pregnancy is of great concern. Our findings, that self-reported snoring is prevalent during pregnancy and therefore increases adverse fetal and birth outcomes is supported by other studies such as O’Brien and Loube.
This study demonstrates an increased need for improved surveillance among the pregnant population. There needs to be a pregnancy specific tool developed for the use of obstetricians and pregnancy care providers which is easy to use and quickly assists in the diagnosis of OSA during pregnancy. Treatment needs to be readily available with rapid access for pregnant women. The role of hypertensive and other disorders (GDM) potentially relating to risk factors for OSA such as overweight/obesity should be explored.
This study has several strengths including the large sample size, variable gestational ages studied, a cross sectional population group including various BMI and ages. Within this study there were multiple racial groups and socioeconomic backgrounds. A prospective design with high completion rates. A limitation of the study is the fact that neither questionnaire used are validated in pregnancy. However, the overall goal of the study was to determine the prevalence of OSA symptoms in pregnancy and to examine their relationship with fetal and obstetric outcomes.
Obstructive sleep apnea is complex multifactorial disease process, superimposed over pregnancy it can become extremely complex. Identification and diagnoses of OSA in pregnancy is a step in the right direction to understanding the implications in pregnancy and the changes to fetal and maternal outcomes. Identification and early treatment may decrease the financial impact on health both short and long term with infants being born vaginally, at the appropriate gestational age with minimal needs for SCN and longer-term medical interventions.
Further studies are needed to determine the prevalence of OSA in the pregnancy compared to age-matched non-pregnant women and to determine whether treating OSA in pregnancy helps improve obstetric outcomes.
Footnotes
Acknowledgements
NH&MRC grant No: 570963
Ethical Approval
The study protocol was approved by the (RNSLHD) Human Research Ethics Committee and ratified at local committees.
Informed Consent
All patients gave written and verbal consent for the research.
Guarantor
The guaranteeing author is Annette Robertson.
Contributorship
All authors have assisted in varying degrees with the following:
• Substantial contributions to the conception or design of the work; or the acquisition, analysis, or interpretation of data for the work; AND
• Drafting the work or revising it critically for important intellectual content; AND
• Final approval of the version to be published; AND
• Agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship and/or publication of this article.
