Over the last three to four decades, it has been observed that the average total hours of sleep have decreased to less than seven hours per person per night. Concomitantly, global figures relating to obesity and diabetes mellitus have increased in an alarming fashion in adults and children, and it has been hypothesized that neuro-hormonal changes accompanying this behavioral sleep deprivation may lead to insulin resistance and, subsequently, to diabetes mellitus. Sleep deprivation has been associated with multiple physiological changes, including increased cortisol and ghrelin levels, decreased leptin levels and impaired glucose metabolism. Experimental studies have also shown an increase in inflammatory and pro-inflammatory markers, which are indicators of body stress, under sleep deprivation. This review elaborates further on this hypothesis, exploring the molecular basis for the link between both entities and the underlying pathophysiology that results in insulin resistance and diabetes mellitus. We review the results of experimental and epidemiological studies, specifically examining the relationship between sleep duration and the immune and endocrine systems.

The field of cellular inflammatory molecules and mediators has expanded over recent decades to explore their serious and significant role in the pathogenesis and modulation of several physiological and pathological conditions. Currently, there is a plethora of clinical and experimental data on the interaction between sleep, inflammatory cascades, neuroendocrine functioning and the immune system [34-36]. New terms have been introduced into this field, such as white adipose tissue, adipokines, and adipose tissue hypoxia, and more terms will be used and described in future research papers as we expand our knowledge. van Leeuwen et al. studied the acute consequences of sleep deprivation and documented increased lymphocyte activation with overproductions of IL-1, IL-6 and IL-17; this group speculated that the same effects persist with long-term sleep deprivation and that this could be linked to an increase in cardiovascular events [37]. Interestingly, a gender difference between sleep duration and IL-6 and CRP was observed [38]. These findings were documented in another study on young, otherwise healthy women in which increased inflammatory markers accompanied disturbed sleep and were thought to be associated with gender specific chronic inflammatory diseases [39]. Several other studies have addressed changes in cytokine levels in response to sleep deprivation. Experimentally reduced sleep hours were found to be associated with increased daytime sleepiness, decreased performance and fatigue [40-44]. In a study of 8 young adults examining the relationship between sleep duration/quality and IL secretion, it was found that acute sleep deprivation increases daytime IL-6 and causes somnolence and fatigue during the following day, whereas post deprivation, night-time IL-6 is decreased and associated with deeper sleep [45]. Another study in 11 subjects suffering from chronic insomnia demonstrated that chronic insomnia is associated with a shift of IL-6 and TNF secretion from night-time to daytime, and this may explain the daytime fatigue and performance decrements associated with this disorder. The daytime shift of IL-6 and TNF secretion, combined with a 24-hour hypersecretion of cortisol (an arousal hormone), may explain insomniacs' daytime fatigue and difficulty in falling asleep [46].

Sleep and feeding are both complex physiological processes requiring complex interactions between internal and external factors and pathways. On the surface, they pass as default human needs that are practiced without thinking on a daily basis since birth. However, exploring the molecular structure controlling both activities gives us an idea about the disorders that occur in abnormal states such as reduced sleep, oversleeping or abnormal feeding behaviors. In recent decades, there has been an exponential increase in our understanding of sleep and feeding physiology. One important epidemiological observation was that of the reciprocal relationship between sleep duration and the prevalence of global obesity over recent decades. Large scale epidemiological data from the United States and other parts of the world have consistently shown an overall increase in body mass index (BMI) over recent decades [69-74]. These figures are alarming and should be taken very seriously by governmental bodies and health care decision makers because obesity is closely linked to a number of serious medical illnesses, such as OSA, diabetes mellitus, hypertension and atherosclerosis, all of which cause huge economic burdens on both a short-term and a long-term basis and are predictors of increased morbidity and premature death [75-77]. Concomitant with the increase in the prevalence of global obesity, another “epidemic” was noticed, manifesting as significant changes in sleep habits with a trend toward sleep deprivation among children and adults. Kripke et al. presented data from a survey that was conducted in the Unites States in the 1960s describing an average sleep duration of 8 to 8.9 hours per night [78]. A few decades later, this figure had dropped to 7 hours per night [79]. More recent data shows a further drop in average levels of night-time sleep in adults to less than 6 hours [80]. In a recent cross-sectional population-based study in the United States, 110,441 adults older than 18 years were studied. Among the recruited subjects, 28.3% reported a sleep duration of ≤ 6 hours, 63.3% reported a sleep duration of 7 or 8 hours and 8.5% reported sleeping ≥ 9 hours [81]. Of note, all of these data were collected from self-reported questionnaires, indicating that they do not represent the actual sleep time. Subjects could be reporting time spent in bed rather than actual sleep duration. To overcome this lack of accuracy, a few studies were designed to use wrist actigraphy to measure total sleep time. One of the first studies to objectively assess sleep duration was conducted by Jean-Louis and colleagues on 273 adult volunteers using wrist actigraphy and showed an average sleep duration of 6.22 hours [82]. A larger group of participants was recruited in the CARDIA sleep study, where 669 adults (age 38-50 years, 58% women and 44% black) were assessed via sleep logs and wrist actigraphy [83]. This study showed interesting findings regarding the role of gender, race and socioeconomic status on overall quality of sleep and total sleep duration. The mean time in bed ranged from 7.8 hours for white women to 7.1 hours for black men, whereas mean sleep duration ranged from 6.7 hours for white women to 5.1 hours for black men. These researchers also found that actigraph-measured sleep duration was shorter than self-reported duration [83]. Redline et al. used polysomnography to assess sleep time in 2685 adults and demonstrated that the average sleep time was 6.1 hours for women and 5.7 hours for men [84]. A large subjective, population-based cohort study was conducted among Chinese adults in Singapore, where 58,044 subjects were recruited from the Singapore Chinese Health Study. Thirty-three percent of eligible subjects reported sleeping 7 hours per night, 27.49% reported sleeping 8 hours per night, 23.43% reported sleeping 6 hours per night, 9.38% reported sleeping less than 5 hours per night and 6.63% reported sleeping more than 9 hours per night [85]. A larger cohort of 104,010 Japanese subjects was analyzed using self-reported data in the Japan Collaborative Cohort Study, and the analysis showed a mean sleep duration of 7.5 hours for men and 7.1 hours for women [86]. Differences obviously exist between self-reported sleep duration and objectively measured sleep duration that are affected by health status, socio-demographic data and sleep characteristics [87].

Several studies have shown a decreasing trend in nocturnal sleep duration among children and young adolescents in various parts of the world [88-90]. Various studies have repeatedly confirmed the side effects of inadequate sleep on growing children and adolescents in particular. Several authors have studied acute and long-term behavioral and neuro-cognitive consequences in school children. Some of these studies were conducted in children and adolescents suffering from concomitant sleep-disordered breathing [91-93]. Fallone et al. examined the effects of acute sleep deprivation on the daytime behavior and performance of healthy children and adolescents. Eighty-two children (age 8 to 15 years) completed 5 nights of baseline sleep and were then randomly assigned to optimal (10 hours) or restricted (4 hours) sleep in the sleep laboratory. Sleep deprivation was associated with shorter daytime sleep latency, increased subjective sleepiness, and increased sleepy and inattentive behaviors [94]. Regarding the metabolic and endocrine consequences of sleep deprivation in children, an early observation of a possible link was made by Locard et al. in a study of obesity risk factors in a five year old population, and short sleep duration was identified as an important risk factor [95]. A Japanese group found a dose-response relationship between short sleeping hours and childhood obesity in a sample of 8274 children (4194 males and 4080 females, age 6 to 7 years). Compared with children who had 10 or more hours of sleep, the adjusted odds ratio was 1.49 for those with 9 to 10 hours sleep, 1.89 for those with 8 to 9 hours sleep and 2.87 for those with <8 hours sleep [96]. Von Kries et al. conducted a cross-sectional study in 6862 German children (age 5 to 6 years) and found a negative correlation between sleep duration and the risk of developing obesity (10 hours, 5.4%; 10.5-11.0 hours, 2.8%; and 11.5 hours, 2.1%). This effect was confirmed after controlling for confounders such as parental obesity, socioeconomic status, and lifestyle factors [97]. Another large scale cross-sectional study was performed in 4511 children (7 to 9 years) and reported an inverse relationship between sleep duration and skin fold thickness as a measure of total body fat. That group identified sleep duration as a modifiable risk factor when managing childhood obesity [98]. The 'Québec en Forme' Project showed similar results in 422 children (5 to 10 years) [99]. Among Arab children, a sample of 5,877 Saudi children (age 10 to 19 years) was studied, and the study group showed that sleeping ≤ 7 hours significantly increased the risk of obesity among boys and girls of all age groups [100]. Obviously, conducting longitudinal studies would be more informative in this regard. However, studies are limited with this design. Agras and his group followed 150 children from birth until the age of 9.5 and found a negative correlation between total sleep hours at the age of 3-5 years and body weight measured at the age of 9.5 years [101]. Taveras et al. conducted an interesting longitudinal study of 915 children who were followed from infancy until the age of 3 years and found that infant sleep of less than 12 hours per day was associated with a higher BMI and obesity [102]. A recent study by Snell et al. evaluated longitudinal data from a nationally representative sample of 2,281 American children (age 3 to 12 years). Baseline assessment of body mass index and body weight was performed at the age of 3 to 5 years and repeated 5 years later. There was a clear association between shorter sleep duration and future increases in body weight [103]. A larger group of children was evaluated for early obesity risk factors in a prospective cohort study in the UK. Short sleep duration was identified among 7 other risk factors predisposing children to childhood obesity [104]. In the longest prospective birth cohort study, 1037 participants born in Dunedin, New Zealand between April 1972 and March 1973 were recruited. Parents reported the sleep durations of the participating children at ages 5, 7, 9, and 11 years, and these data were analyzed in relation to the participants’ BMIs at the age of 32 year. Interestingly, shorter childhood sleep times were significantly associated with higher adult body weights, and this association was present after adjusting for early childhood weight, socioeconomic status, parental BMIs and lifestyle [105]. Touchette at al. also conducted a longitudinal study of 1138 children where BMI was measured at ages 2.5 and 6 years. Sleep duration was reported yearly by their mothers. The study group identified 4 sleep duration patterns: short persistent (sleeping < 10 hours per night: 5.2%), short increasing (sleeping less than 10 hours per night at the age of 29 months to more than 10 hours per night at 41 months and maintaining this level until 74 months of age: 4.7%), 10 hour persistent (sleeping persistently 10 hours per night: 50.7%), and 11 hour persistent (sleeping 11 hours per night: 39.4%). Persistently short sleep durations (<10 hours) during early childhood significantly increased the risk of obesity by a factor of 4.2 compared to 11 hour persistent sleepers [106]. Wrist actigraphy was used to measure sleep duration in 591 children in New Zealand, and the results have been published recently. Mean time in bed according to parental reports was 10.9 hours, and mean sleep duration measured using actigraphy was 10.1 hours. In a multivariable analysis, sleep duration varied according to weekday, season, and the presence of younger siblings. Again, shorter sleep duration was found to be an independent risk factor for higher body weight in children [107]. Finally, a systematic review and meta-analysis of published data in this field, mainly from cross-sectional studies, was recently published by Chen et al. This review drew the following conclusions: 1) children with shorter sleep durations had a 58% higher risk for obesity, and children with the shortest sleep durations had an even greater risk (92%) than children who had longer sleep durations. This risk is reduced by 9% for each hour increase in sleep duration. 2) There is a significant linear dose-response relationship, but only in young children. 3) There is a significant gender difference (with a greater risk in boys) in the trend toward the development of obesity with shorter sleep durations [108].

With regard to data concerning endocrine and metabolic consequences of voluntary or behavioral sleep deprivation in adults, we will review epidemiological and experimental studies in this field. Obesity is an important risk factor in the pathogenesis of various diseases such as diabetes mellitus, OSA and metabolic syndrome. More importantly, obesity was found to decrease life expectancy by 7 years at the age of 40 [76]. Obesity is known to be genetically and environmentally driven. One of the recently identified risk factors is a change in sleep quality or quantity. In recent decades, several studies have been conducted to determine whether there is a cause-effect relationship between sleep duration and body weight, and the data indicate that interpersonal differences in the sensitivity to sleep loss exist [109]. A cross-sectional study was conducted in 983 community-dwelling elderly subjects (age 57 to 97 years), among whom sleep duration was assessed using wrist actigraphy. The resulting data showed a significant U-shaped relationship with BMI, and both short sleepers (<5 hours) and long sleepers >8 hours) were more likely to be obese [110]. Watanabe and his group conducted a prospective study in a large sample of Japanese employees (35,247) with a 1 year follow-up to investigate the association between short sleep duration and elevated BMI and obesity. Short sleep duration was associated with weight gain and the development of obesity over 1 year in men, but not in women [111]. In a cross-sectional study that objectively assessed the association between sleep duration and body weight in elderly people, wrist actigraphy was performed in 3055 men (67-96 years) and 3052 women (age 70 to 99 years), and polysomnography was performed in a subgroup of 2862 men and 455 women. Compared to those sleeping an average of 7 to 8 hours per night, sleep duration of less than 5 hours was associated with a BMI that was on average 2.5 kg/m² greater in men and 1.8 kg/m² greater in women [112]. The same group also investigated the association between self-reported usual sleep duration and subsequent weight gain in the 68,183 women who participated in the Nurses' Health Study. It was found that women sleeping 5 hours or less gained 1.14 kg more than those who slept an average of 7 hours over 16 years, and women sleeping an average of 6 hours gained 0.71 kg more. The relative risks ratio of a 15-kg weight gain were 1.32 and 1.12 for those sleeping 5 and 6 hours, respectively [113]. Data from the Quebec Family Study, which represents a 6 year observational study in 276 adults (age 21 to 64 years), provided evidence that long sleeping times predict an increased risk of future body weight and adipose tissue gains in adults [99]. López-García and her colleagues examined the association of habitual sleep duration with obesity and weight changes among 3576 persons aged 60 years or more in Spain. Subjects who slept 5 hours had a greater frequency of obesity with an odds ratio of 1.33 and a greater frequency of severe obesity with an odds ratio of 2.08 than subjects who slept 7 hours [114]. Vorona et al. examined the association between restricted sleep and obesity in a heterogeneous adult primary care population (924 patients, age 18 to 91 years). They found that the mean BMI was 30 kg/m², women slept more than men, and overweight and obese patients slept less than patients with a normal BMI [115]. Shigeta at al. found in 453 subjects (321 men and 131 women) that sleeping less than 6 hours was associated with an increased risk for the development of obesity [116]. A Korean group examined the association between sleep duration and general and abdominal obesity in 8,717 adults (age 20 to 65 years) from the Korean National Health and Nutrition Examination Survey (2001-2005). After controlling for socio-demographic and lifestyle factors, the adjusted odd ratios associated with sleeping 5 hours or less per day versus 7 hours per day were 1.25 for general obesity and 1.24 for abdominal obesity [117]. Hasler et al. tested the hypothesis that short sleep duration leads to weight gain during early adulthood [118]. This was a prospective single-age cohort study of 496 young adults with 4 interviews conducted at the ages of 27, 29, 34, and 40 years. There was a clear association between short sleep duration and obesity at the age of 27 years. However, this association diminished after the age of 34 years [118]. Finally, in a systematic review of the current literature regarding short sleep duration as an independent risk factor for obesity and weight gain, the results from cross-sectional studies in adults were mixed. In contrast, longitudinal studies showed a positive association between short sleep duration and future weight gain [109]. In the second meta-analysis of data in this field, Cappuccio et al. pooled representative data from 26 studies including 604,509 adults worldwide and found a consistently increased risk of obesity among short sleepers in adults [119].

Sackner et al. in 1975 described patients with OSA as being chronically sleep deprived and noted that this condition was linked to the development of systemic arterial hypertension [120]. Ten years later, Tochikubo and his group monitored blood pressure in a group of young healthy males (age 23 to 48 years) who were experimentally sleep-deprived to a mean duration of 3.6 hours of sleep. Blood pressure, heart rate and urinary excretion of norepinephrine were all elevated on the following day [121]. Another group studied blood pressure, heart rate, forearm vascular resistance, and muscle sympathetic nerve activity measured at rest and during four stressors. These measures were obtained twice, initially after normal sleep and later after a night of sleep deprivation in 8 healthy subjects (age 35 to 45 years, six men and two women). Sleep deprivation resulted in an increase in blood pressure and a decrease in muscle sympathetic nerve activity [122]. Heart rate, forearm vascular resistance, and plasma catecholamines were not significantly changed by sleep deprivation, nor did sleep deprivation affect autonomic and hemodynamic responses to stressful stimuli [122]. In a longitudinal study (4810 subjects) that assessed the effect of short sleep duration on blood pressure and the risk for hypertension, Gangwisch et al. followed participants for 10 years, and hypertension was diagnosed in 647 subjects [123]. A sleep duration of 5 hours per night was associated with a significantly increased risk of hypertension, even after controlling for other risk factors [123]. A study of 6 young, healthy men who were acutely sleep-deprived showed that diastolic blood pressure was significantly higher after total sleep deprivation than after normal levels of sleep [124]. An interesting study was conducted by Irwin and Ziegler of 36 abstinent, alcohol-dependent men and 36 matched controls with similar baseline parameters [125]. Partial night sleep deprivation induced greater increases in heart rate and circulating levels of catecholamines in the alcohol-dependent men than in the controls [125]. Stranges et al. examined cross-sectional and prospective associations of sleep duration with prevalent and incident hypertension in a cohort of 10,308 British civil servants (age 35 to 55 years) and found a gender difference, where women were at a higher risk for developing hypertension than men [19]. A similar gender difference was found by Stang et al. when examining a cohort of subjects from the Heinz Nixdorf Recall Study, and sleeping for less than 5 hours per night was found to be a risk factor for hypertension in women [126]. Vgontzas et al. reported an association between sleep duration and hypertension in 1,741 men and women in a cross-sectional study. The risk was highest in subjects who slept less than 5 hours per night and in those with poor sleep quality [127]. A recent study in a group of 238 healthy adolescents used multiple-day wrist actigraphy at home and in-lab polysomnography for accurate sleep duration assessments [128]. Again, both short sleep duration and poor sleep quality were associated with a higher propensity for the development of hypertension [128].

Recent data suggest an association between sleep deprivation and dyslipidemia. Kaneita at al. examined the individual association between sleep duration and high serum triglycerides, low HDL (high-density lipoproteins) cholesterol, or high LDL (low-density lipoproteins) cholesterol levels. The study analyzed data from the National Health and Nutrition Survey including 1,666 men and 2,329 women aged 20 years or older and showed that among women, sleeping less than 5 hours per night or more than 8 hours per night were associated with high serum triglyceride levels or low HDL cholesterol levels [129]. Importantly, this finding was persistent even after adjusting for confounders [129]. The Hordaland Health Study evaluated sleep habits in 8860 subjects (aged 40 to 45 years) and compared them to body weight, blood pressure, total cholesterol, HDL cholesterol and triglyceride levels. Short sleep durations were associated with elevated BMIs, higher non-fasting levels of cholesterol and triglycerides, and higher systolic and diastolic blood pressures [20]. A recent study of a large group of young adults ( 14,257 subjects) who were followed from 1994 to 2002 showed that short sleep durations could be a significant risk factor for hypercholesterolemia in young adults and, in particular, females [123]. The authors suggested that lengthening sleep could potentially both treat and act a primary preventative measure for hypercholesterolemia. The Rotterdam Study, on the other hand, used wrist actigraphy to assess sleep duration in an older group of subjects (768 subjects, age 57 to 97 years) [130]. Sleep duration and sleep quality were associated with higher cholesterol levels [130].

Sleep and eating are vital physiological processes that are linked via a complicated network of hormonal pathways, and losing the homeostasis of one process exerts deleterious effects on the other during both the short and the long term. The link between over- and under-sleeping and glucose dysregulation is a very serious breakthrough in our understating of obesity, diabetes mellitus and metabolic syndrome pathophysiology, and this should be addressed more methodically by epidemiologists and researchers to identify optimal sleeping hours. For the time being, no strong recommendations can be made as to how many hours of restorative sleep we actually need to maintain a healthy life and normal body weight and to avoid the development of the above-mentioned serious metabolic consequences.