The Relationship and Potential Mechanistic Pathways Between Sleep Disturbances and Maternal Hyperglycemia

Abstract

This paper reviews recent work investigating the influence of sleep disturbances on maternal hyperglycemia, particularly gestational diabetes mellitus (GDM). The incidence and prevalence of hyperglycemia are increasing worldwide, which is cause for concern because GDM and even mild hyperglycemia are associated with adverse pregnancy outcomes. A better understanding of sleep-related risk factors for maternal hyperglycemia is an important health matter. Evidence demonstrates associations between sleep disturbances, especially sleep-disordered breathing, and hyperglycemia, but causal effects and the underlying mechanisms linking these conditions have not been fully elucidated. Subjective sleep assessments show associations between sleep disturbances and maternal hyperglycemia. There are, however, few studies using objective measures to support these findings. Large prospective studies are required to examine causal relationships between sleep disturbances and maternal hyperglycemia. There is also a need for smaller mechanistic studies to understand the pathophysiology. Furthermore, interventional studies are required to address whether improvement of sleep parameters can prevent/decrease the risk of developing maternal hyperglycemia. Taken together, the data suggests that sleep disturbances during pregnancy are important to identify and manage in order to minimize maternal hyperglycemia and GDM, and improve maternal and fetal well-being.

Introduction

Hyperglycemia occurring during pregnancy refers to a continuum of disease severity ranging from glucose concentrations slightly above normal to gestational diabetes mellitus (GDM) [1-4]. GDM complicates approximately 7% of all pregnancies, affecting 200,000 pregnant women annually in the US [5]. The prevalence rises with increasing rates of maternal obesity and type 2 diabetes mellitus [2 4].

GDM and mild hyperglycemia even in non-diabetic ranges are associated with adverse maternal, fetal and neonatal outcomes [3 4]. Mothers with GDM have an increased risk of pre-eclampsia, cesarean section, preterm labor from polyhydramnios and infection [3 6 7]. GDM is also a risk factor for type 2 diabetes, obesity, cardiovascular disease and metabolic syndrome among both mothers and infants [4 8]. Furthermore, infants of diabetic mothers are at increased risk of macrosomia, related birth injuries, respiratory problems, neonatal hypoglycemia, jaundice [4 6] and impaired intellectual and psychomotor development [9]. GDM increased national medical costs by $636 million in 2007—$596 million for maternal costs and $40 million for neonatal costs in the US [10]. The causes of GDM and its rising incidence remain uncertain, though it is presumed to be related to obesity. Therefore, identifying modifiable risk factors that might help prevent GDM is of urgent public health importance.

Emerging literature shows that sleep disturbances during pregnancy are associated with hyperglycemia. In this article, sleep disturbances refer to internal and external factors that change the duration and/or structure of normal sleep architecture and cause poor sleep quality and daytime sleepiness, including disorders of initiating and maintaining sleep, disorders of the sleep–wake schedule, and dysfunctions associated with sleep, sleep stages, or partial arousals (transitions from deep sleep stages to lighter stages and/or partial wakefulness) in the absence of mental disorders, prescribed medications, substance abuse, or medical conditions [11]. We will review recent data examining associations between GDM and sleep disorders and sleep parameters in pregnant women, focusing especially on sleep quality, sleep duration and sleep-disordered breathing (SDB). We will also examine potential underlying mechanisms for these relationships; identify gaps in scientific knowledge; and propose directions for future research.

Sleep characteristics and disturbances during pregnancy

Sleep disturbances are surprisingly common during pregnancy. The vast majority of pregnant women (over 75%) experience sleep difficulties in the form of multiple nocturnal awakenings [12], likely due to hormonal changes, physical discomfort and anxiety associated with fear of labor or having an unhealthy baby and forthcoming life changes [13 14]. Sleep disturbances can commence with the onset of pregnancy (11 to 12 days), and increase in frequency as pregnancy progresses [15-19]. Some complaints associated with sleep disturbances (Table 1) are specific to one trimester, such as vomiting (morning sickness), but others occur across trimesters, such as urinary frequency and leg cramps [15 16 19-23]. A longitudinal survey study of 325 Finnish women showed that the number of nocturnal awakenings was highest in the third trimester [15]. A recent prospective study of 260 women using the Pittsburgh Sleep Quality Index (PSQI) [24] reported that sleep quality deteriorated from the second to the third trimester. Thirty-six percent of women were poor sleepers in the second trimester and 56% of women in the third. A third study using polysomnography (PSG) also found more nocturnal awakenings and lower sleep efficiency in late pregnancy compared to early pregnancy and the non-pregnant state [25].

Table 1.

The most common etiologies of sleep disturbances during pregnancy

Sleep Disturbances	1st Trimester	2nd Trimester	3rd Trimester
Nighttime
General discomfort	+	+	+
Psychological Stress	+	+	+
Urinary frequency	+	+	+
Low back pain/joint pain	+	+	+
Leg cramps	+	+	+
Heartburn		+	+
Snoring		+	+
Dreams and nightmares		+	+
Fetal movements		+	+
Irregular uterus contractions		+	+
Nausea and vomiting	+
Daytime
Fatigue	+		+
Daytime sleepiness	+		+

Sleepiness and fatigue are the most common first trimester complaints. Sleepiness has been attributed to increased progesterone levels because of its sedative effect and inhibitory effect on muscles, which may result in frequent urination and snoring in late pregnancy [21]. In the first trimester of pregnancy, sleep has a polyphasic character including nocturnal sleep and often one or more daytime naps. Although the majority of women (75%) reported napping during pregnancy, some may not [12]. In the second trimester, sleep again becomes monophasic as in non-pregnant women [26]. By late in the second trimester, total nocturnal sleep time decreases. Total sleep time tends to be longer in the third trimester than pre-pregnancy period due to frequent daytime naps but sleep efficiency and percentage of slow wave sleep (SWS: deep sleep, consisting of stages 3 and 4 of Non-REM, occurring early in the sleep cycle) are lower than in the pre-pregnancy period or in early pregnancy [15 19 23 25]. The physiologic and biochemical changes of pregnancy may place women at risk for developing specific sleep disorders such as obstructive sleep apnea (OSA) and restless legs syndrome (RLS) [16 27-29]. Pregnancy may also exacerbate existing sleep disorders [19 23 30 31]. However, sleep disorders during pregnancy are frequently undiagnosed or overlooked.

Why are sleep disturbances serious during pregnancy?

Maternal sleep quality and duration can affect a number of endocrine, metabolic, and neurological functions that are critical to maintain a healthy pregnancy and fetal growth [21]. Thus sleep disturbances may be detrimental to fetal growth and maternal well-being. The collective evidence from epidemiological, experimental and clinical studies in the general population demonstrates associations between sleep disturbances and an array of adverse health conditions such as hyperglycemia, insulin resistance, type 2 diabetes, metabolic syndrome, coronary artery disease and stroke [32-37]. Emerging evidence suggests an analogous association in pregnancy such that women with SDB and short sleep duration (SSD) as assessed by questionnaires, actigraphy or PSG are at increased risk of developing adverse maternal and fetal outcomes such as GDM and pre-eclampsia [18 20• 38-41•••]. Several studies have reported that SSD increased the duration of, and pain perception during, labor, with a higher rate of caesarean delivery and preterm birth [42 43]. Although the results of these studies should be interpreted cautiously due to their small sample sizes, recall bias and other study limitations, this evidence supports the concept that extremes of sleep duration and sleep disorders such as SDB can precipitate or exacerbate a number of pregnancy complications, including GDM.

How obesity, sleep and glucose metabolism are related in pregnancy/postpartum

Obesity has been regularly reported to be associated with increased risk for pregnancy complications, including pre-eclampsia, hypertension, and GDM. The prevalence of obesity has been increasing steadily among reproductive-aged women in many parts of the world, with rates ranging from 10% up to 38% in Western societies [2 44 45]. Existing data shows that obesity is a major risk factor for insulin resistance, GDM and SDB [4 6 40•• 46], a term that refers to a spectrum of abnormal respiratory events during sleep ranging from habitual snoring to OSA, in which recurrent episodes of upper airway collapse occur during sleep [21]. This section will examine the associations between obesity, sleep disturbances (especially SDB) and maternal hyperglycemia, conditions that can potentiate each other, leading to a vicious cycle. Women who are obese prior to or during pregnancy are more likely to suffer from SDB than lean women [40•• 47]. Pregnant women have decreased lung reserve due to pregnancy-related changes [48 49]. The combination of obesity and pregnancy may have adverse effects on the respiratory system, leading to SDB. Conversely, SDB may predispose women to worsening obesity because of sleep fragmentation, daytime sleepiness, and insulin resistance. As reviewed below, SDB may also be associated with changes in leptin and ghrelin levels, increased appetite and caloric intake, again exacerbating obesity, which potentially contributes to hyperglycemia.

Human and animal studies have reported that short or long sleep duration (generally defined as <6 h and >9 h, respectively), sleep restriction and sleep disturbance have metabolic effects that predispose to weight gain and impaired glucose metabolism [32 33 50 51]. Although there is limited evidence regarding whether a relationship exists between sleep duration and weight gain in pregnancy, a number of studies have examined associations between postpartum sleep duration and weight retention. A prospective study of 904 postpartum women found that women reporting short sleep duration (<5 hours/day) at 6 months postpartum were 2.3 times more likely to retain at least 5 kg at 1 year independent of potential confounders including pre-pregnancy body mass index (BMI), gestational weight gain, parity, and postpartum behaviors [52]. Another prospective cohort of 586 women reported that postpartum sleep ≤5 hours/day was associated with higher postpartum weight retention, higher subscapular + triceps skinfold thickness and higher waist circumference at 3 years postpartum, after adjusting for age, pre-pregnancy BMI, and other relevant confounding factors [53].

Although poor sleep quality and sleep loss among pregnant women can occur as a result of hormonal and physical factors [1], voluntary sleep restriction is also common due to work and household responsibilities. Sleep loss or self-imposed SSD during pregnancy may also impair glucose regulation given that women who sleep less have more opportunities to eat. Furthermore, sleepiness and fatigue, which are common conditions in pregnancy, are independently associated with insulin resistance [54•], resulting in increased risk of obesity [2]. Stress and fatigue due to inadequate sleep may also lead to comfort eating, physical inactivity and eventually obesity. Thus, avoiding obesity before and during pregnancy may minimize the risk of developing sleep disorders and hyperglycemia during pregnancy.

Does lack of sleep cause maternal hyperglycemia during pregnancy?

In non-pregnant populations, SSD and fragmented sleep are known as major determinants of metabolic health, and are associated with poor glucose control in patients with diabetes [33 34 37 55-57]. Some prospective studies have also reported that long sleep duration and poor sleep quality and sleep maintenance [32-34 58] are associated with a higher risk of impaired glucose tolerance or type 2 diabetes.

In the pregnant population, data in this area are conflicting. The majority of studies demonstrate an association between sleep duration and increased risk of hyperglycemia [18 50 59•]. In a prospective study of 189 nulliparous women with a singleton gestation, 26% of women reported SSD (<7 hours sleep/night) in the first trimester, 40% in the third trimester and 48% of all women in one or both trimesters. Women who reported SSD were 2.6 (1.3-5.7) times more likely to have 1-hour oral glucose testing values ≥130 and GDM (OR 10.6, 1.3-85.5) after adjusting for age, race, prepregnancy BMI and snoring [18]. In a large prospective cohort study of 1290 women before 20 weeks gestation, self-reported sleep duration was categorized as ≤4, 5-8, 9 (the reference sleep category), or ≥10 hours/night. Regardless of maternal pre-pregnancy weight, both short and long sleep durations were associated with higher frequencies of glucose intolerance (1 hr oral glucose tolerance test (OGTT) ≥140 mg/dl). Women sleeping ≤4 hours/night had a 5.56-fold (1.31-23.69) increased risk of GDM compared with women sleeping 9 hours/night after adjusting for age and race. However, the association was no longer significant after adjusting for pre-pregnancy BMI [50].

In a third study of 169 women who completed several sleep questionnaires, including Epworth sleepiness scale (ESS) and PSQI. There was a significant inverse correlation between sleep duration and 1-h glucose values following 50-g OGTT such that one hour of shorter sleep was associated with a 4% increase in glucose levels. Furthermore, short sleep and measures of poor sleep quality were also correlated with preterm delivery and greater rate of admission to neonatal intensive care units in both GDM and non-GDM women. However, in this study age, pre-pregnancy BMI, a family history of diabetes and previous GDM were not controlled [59•]. A prospective study of 260 pregnant women using the PSQI reported that sleep quality worsened during pregnancy, but sleep parameters were not significantly different between pregnancies with and without adverse outcomes including GDM [24].

Studies with relatively small sample sizes (n=52-104) have tended not to find associations between sleep duration and glucose intolerance or GDM [20• 60•]. In a secondary analysis of a prospective study of 104 pregnant women, neither objective nor subjective sleep duration was related to hyperglycemia after controlling for known risk factors for GDM [20•]. However, self-reported nap duration was associated with high glucose challenge test (GCT) results in adjusted models. Similar results were reported by a recent case-control study (n=52) that found no significant differences in either objective or subjective measures of sleep duration between women with and without GDM [60•].

Findings showing an association between longer self-reported nap duration in early pregnancy and high GCT [20•] are compatible with results from studies of older adults, among whom daytime napping has been associated with a higher risk of diabetes [61 62]. Results from Bourjeily [54•] et al. also have demonstrated that severe daytime sleepiness increases the risk of GDM in nonsnoring pregnant women, independent of confounders. These studies suggest that sleep has modulatory influences on glucose regulation regardless of time of day. Hypothetically, disturbed sleep/wake cycles due to napping at unusual times may disturb the sympathetic-parasympathetic balance and cause excessive hormone release that regulate glucose metabolism [20•]. Daytime naps also reduce the amount of SWS, which is associated with glucose homeostasis as detailed below [63]. To elucidate whether napping has clinical implications for the treatment and prevention of GDM, further studies are required.

Is OSA an independent risk factor for glucose dysregulation?

The prevalence of OSA, diagnosed by PSG, in pregnant women is not known because large population-based epidemiological studies using objective sleep measures are lacking. However, self-reported snoring and daytime sleepiness, two symptoms of OSA, are very common. The prevalence of habitual snoring among pregnant women has been estimated to be between 14% and 46% [16-18 28 39• 46 47 50 59• 64]. Longitudinal studies have shown that frequent snoring (≥3 nights/week) increases as pregnancy progress [17-19 39•]. The combination of weight gain and pregnancy-induced physiological, anatomical and hormonal changes in the respiratory system (upper airway, lung mechanics, and control of breathing) make women more prone to snoring and increase the risk of developing OSA and nocturnal desaturation. SDB may be particularly severe during the third trimester [16 28 39•] when oxygen stores in the lung are reduced due to lung compression from the enlarging uterus.

In the pregnant population, most observational studies have demonstrated an association between SDB and hyperglycemia independent of confounders including race and BMI. A prospective cohort study of 1290 pregnant women reported that 1-hour OGT values were higher among women with habitual snoring after adjusting for age, race/ethnicity and BMI (p=0.04). Overweight or obese women who snored had a significantly increased risk of GDM compared to lean nonsnorers (RR 6.91, 95% CI 2.87-16.6). However, the odds of GDM in all snorers were not significantly increased compared to nonsnorers (RR 1.86, CI 0.88-3.94) [50]. Similarly, a large cross-sectional survey of 1000 pregnant women found an association between symptoms of SDB and GDM after adjusting for multiple factors including BMI at delivery (2.1, 95% CI 1.3–3.4) [64]. In a prospective study, 189 healthy nulliparous women completed a sleep survey both at study enrollment and in the third trimester. This study showed that women who reported frequent snoring at both time points were 4.9 times (95% CI, 1.3–18.1) more likely to develop GDM after controlling for BMI compared to nonsnorers [18]. Similarly, in a prospective study of 169 women, a significantly higher risk of developing GDM was observed among pregnant women at increased risk for SDB (after adjustment for BMI) or who endorsed a combination of increased SDB risk and short sleep duration [59•]. Balserak et al also found that self-reported symptoms of SDB in early gestation were independently associated with high GCT values (≥135) in a secondary analysis of a prospective study [20•].

In contrast to these studies, a large, prospective study of 1719 women in the 3rd trimester in which a sleep survey was administered to a clinic-based population demonstrated that neither pregnancy-onset nor chronic snoring was associated with GDM after adjusting for confounders [39•]. Although self-reported SDB symptoms have been linked with hyperglycemia in most of these studies, one must be cautious when interpreting these results because SDB questionnaires have low sensitivity and specificity for predicting OSA in pregnant women or have not been validated in pregnancy.

Studies objectively assessing the impact of SDB on the risk of hyperglycemia in pregnant women are limited and the results of these studies are contradictory. Chen et al. [41•••] convincingly reported that the risk of GDM increased almost two-fold (OR 1.6, 95% CI 1.07–2.8) among women with PSG-confirmed SDB in a large retrospective cohort study involving 759 pregnant women with OSA and 3955 matched controls. A recent retrospective cohort study of 143 women with confirmed moderate to severe OSA (AHI≥15) also demonstrated an increased risk of composite adverse pregnancy outcomes (GDM and pregnancy related hypertension), after adjusted for timing of the sleep study performed before, during or after pregnancy. However, the effect of obesity was not fully controlled because BMI used in the statistical analysis was ascertained at the time of the sleep study [65•]. Negative associations between objectively measured AHI and hyperglycemia have reported in studies with smaller sample sizes (n=52-104) [20• 60•]. For instance, in a cohort study of 175 obese pregnant women, OSA was associated with increased prevalence of pre-eclampsia, neonatal intensive care unit admissions, and cesarean delivery, but the prevalence of GDM was not increased [40••]. This may be due to the study's relatively small sample size and because all subjects were obese.

In the Sleep Heart Health Study, individuals with diabetes had an increased risk of poor sleep and some form of SDB (up to 58%), which may impair glucose control [55]. These findings suggest that women with pre-pregnancy diabetes should be screened for OSA [21]. Intervention studies have shown that treatment for OSA with CPAP improves glucose control in non-gravid populations [34 66]. However, whether OSA intervention improves outcomes in pregnancy and whether it may prevent pregnancy complications associated with GDM in the long term is not known. Because of the unique nature of pregnancy, treatment efforts for OSA need to address the distinct attributes of intervening in pregnant women (e.g. frequent nocturnal urination), including issues of CPAP initiation and adherence.

Restless legs syndrome and glucose metabolism

Restless leg syndrome (RLS) is characterized by an urge to move the extremities in response to uncomfortable sensations, a circadian pattern in which symptoms worsen at night, and are relieved by movement. RLS leads to difficulties in initiating and maintaining sleep and affects approximately 25% of women during pregnancy [21 22 29]. RLS has been shown to have a significant association with type 2 diabetes in the non-pregnant population [67 68]. In the pregnant population, little data exists, with only a recent preliminary study showing that women with GDM are significantly more likely to have RLS than age and BMI matched women [60•].

Narcolepsy and glucose metabolism

Narcolepsy is characterized by excessive daytime sleepiness, various combinations of irresistible onset of sleep, REM-related sleep symptoms (e.g. sleep paralysis and hypnagogic or hypnopompic hallucinations), and cataplexy. Cataplexy however is not present in all narcoleptics [69]. Honda et al found that narcolepsy was associated with increased frequency of type 2 diabetes [70]. However, subsequent studies have failed to find evidence that narcolepsy increases the risk of insulin resistance or glucose intolerance (and consequently of type 2 diabetes) independently of BMI [71 72].

In a retrospective cohort study (n=54), pregnant women whose narcolepsy symptoms started before or during pregnancy tended to have greater glucose intolerance or more severe type 2 diabetes mellitus compared to the asymptomatic group. The factors associated with glucose metabolism such as weight gain during pregnancy did not differ between these groups [69]. A recent retrospective cohort study undertaken in European countries also reported that the symptomatic patients with narcolepsy-cataplexy had a significantly higher prevalence of impaired glucose metabolism than the asymptomatic patients [73]. However, these studies had no control groups without narcolepsy.

What are the potential mechanisms underlying the link between sleep disturbances and hyperglycemia?

Mechanisms linking sleep disturbances and pregnancy complications including hyperglycemia are likely multifactorial and are not completely understood. Both Type 2 diabetes and GDM are associated with insulin resistance and impaired insulin secretion, and share the same risk factors and genetic susceptibility [74]. However, normal pregnant women are mildly insulin resistant due to hormonal changes [2 4 74]. Thus, even small changes in sleep parameters could potentially make pregnant women more vulnerable to developing maternal hyperglycemia.

Figure 1 provides guidance for our proposed model explaining potential pathways in pregnancy. Our model suggests that habitual poor sleep quality or SSD, specifically insufficient SWS, can lead to impaired glucose tolerance. As has been described in sleep deprivation studies, SSD and decreased SWS may augment the inflammatory response by increasing circulating concentrations of interleukin-6 (IL-6), tumor necrosis factor alpha (TNFa) and C-reactive protein (CRP) [75-80], which are involved in the pathogenesis of insulin resistance and type 2 diabetes [1]. Decreased PSG sleep duration was also associated with increased TNFa levels in the Cleveland Family Study (n=614); this study reported that increases, but not decreases, in habitual sleep duration were linked to elevated CRP and IL-6 levels [81]. Similarly, in mid and late pregnancy, SSD and poor sleep quality have been found to be associated with higher levels of IL-6 [82]. Naps during pregnancy could also activate pro-inflammatory pathways, leading to hyperglycemia. Increased levels of cytokines can decrease insulin sensitivity and downstream insulin signaling, adversely impacting glucose regulation [2 83].

Figure 1.

Schematic illustration of potential causal pathways that may link sleep / sleep disturbances during pregnancy and hyperglycemia.

SWS may have more significant implications for glucose homeostasis than sleep duration alone due to its important role in transitory metabolic, hormonal, and neurophysiologic changes. These include decreasing brain glucose utilization and sympathetic activity, stimulating growth hormone release, inhibiting the activation of the HPA axis and cortisol secretion [35 84 85], all of which likely affect glucose regulation. Experimental studies in the non-pregnant population have observed that reduction in SWS is associated with increased risk of impaired glucose tolerance and insulin sensitivity [34 35 84]. The duration of SWS decreases in most pregnant women by the 3rd trimester [19 23 25]; we believe that a greater decline in SWS is likely to occur among those who develop SDB.

Our model also posits that recurrent partial or complete airway obstruction results in intermittent hypoxia and frequent arousals from sleep in pregnant women. In the general population, arousals and hypoxia decrease total sleep time in general and SWS in particular [86], and also increase sympathetic activity [1 34 84 87] during sleep, which carries over into wakefulness [35]. Furthermore, intermittent hypoxia results in activation of the hypothalamic-pituitary axis (HPA) and alterations in inflammatory pathways [87 88]. Analogous biological pathways could be involved in maternal hyperglycemia.

Intermittent hypoxia-reoxygenation can initiate oxidative stress, creating a vicious cycle that activates inflammatory pathways [87 89]. Alternately, hyperglycemia has been found to acutely increase circulating cytokine concentrations through an oxidative stress mechanism [87 90]. Both oxidative stress and inflammation are linked to endothelial dysfunction [91], which is present in individuals with GDM [83] and in both obese and non-obese women with a history of GDM, even when they have normal glucose tolerance [92]. These findings suggest that endothelial dysfunction may play a role in the onset of GDM.

Sleep fragmentation, proinflammatory cytokines and intermittent hypoxia can also lead to hyperactivation of the HPA, which in turn leads to increased release of the glucocorticoid cortisol [84 88 93 94]. Cortisol regulates several basal processes including fat and glucose metabolism, blood pressure, inflammatory and immune responses. Long-term secretion of the glucocorticoid cortisol increases susceptibility to insulin resistance and impaired glucose tolerance, and restrains further development of the inflammatory process [84 93 94]. Mean 24-hour plasma cortisol levels have been found to be significantly higher in short sleepers than long sleepers [88]. Disruptions in sleep as a result of emotional stress associated with pregnancy could also increase cortisol levels due to activation of the HPA axis [95 96]. Other causal routes, however, may exist. For instance, stressed women may smoke or eat more which contribute to SDB; stress due to poor or inadequate sleep can also affect immune function and vasculature [96]. Further research is needed to better understand the pathophysiology of maternal hyperglycemia regarding the functional relationships between sleep, stress and HPA axis activity.

Recent evidence also suggests that sleep loss, especially loss of SWS, may affect hormones that regulate appetite and satiety, specifically by reducing leptin sensitivity and increasing levels of ghrelin [33 85]. Among police officers, a U-shaped association was observed. Leptin levels were higher at both short and long extremes of sleep duration. These associations were stronger among women [97]. Leptin levels also significantly correlated with insulin resistance in patients with moderate-to-severe OSA. After 8 weeks of CPAP treatment, insulin secretion capacity improved and leptin levels fell [66].

In pregnancy, either long sleep duration (e.g. due to daytime naps) or sleep loss caused by sleep restriction or sleep problems could contribute to the pathophysiology of GDM [18 20• 41••• 50 59• 64 98], by prompting food intake, weight gain and increased insulin resistance [99 100]. However, data on leptin in GDM are controversial. While some studies have reported that maternal leptin concentrations increased in GDM [4, 98, 99, 101] and hyperleptinaemia in early pregnancy was a predictor of developing GDM in later pregnancy, other studies have found that leptin concentrations were unchanged or decreased among women with GDM [99]. Ghrelin mRNA expression in placental tissue has also been found to be higher in individuals with GDM than in controls, whereas there was no association between circulating ghrelin and GDM [98]. In the pregnant population, all the potential pathways linking sleep and maternal hyperglycemia mentioned above are, due to our lack of knowledge about them, important topics of research that warrant further investigation.

Conclusion

The majority of pregnant women experience sleep disturbances due to pregnancy-related changes that can result in poor sleep quality, insufficient sleep and daytime sleepiness. A burgeoning collection of data demonstrates that extremes of sleep duration, poor sleep quality, severe daytime sleepiness and sleep disorders including SDB and symptomatic narcolepsy during pregnancy are associated with maternal hyperglycemia and GDM. Obesity is a major risk factor for these conditions. Therefore, studies investigating sleep disturbances, especially SDB and GDM, in pregnancy must consider obesity as a significant confounder. The physiological stress of sleep disturbances appears to trigger a cascade of pathophysiological events including oxidative stress, sympathetic over-activity, inflammation and endothelial dysfunction that increase the predisposition for insulin resistance, glucose intolerance and GDM. Maintaining a proper sleep pattern and avoiding obesity are critical for glucose regulation and healthy metabolic function in pregnancy. Factors that would improve future investigations in this area include the use of interventions, objective as well as subjective sleep measurements, and utilization of a large longitudinal study design.