Problems Associated with Short Sleep: Bridging the Gap between Laboratory and Epidemiological Studies

Abstract

Existing data from laboratory studies suggest a number of negative consequences of acute reductions in sleep time. Also, epidemiological data suggest links between shorter self-reported sleep duration and negative health outcomes. These bodies of work are growing, revealing several key points of convergence and opportunities for future exploration. In addition, they begin to highlight possible problems experienced by “short sleepers,” who sleep approximately six hours or less per night. While it is likely that this group is heterogeneous, comprised both of individuals with less need for sleep and those not sleeping enough, the laboratory and epidemiological findings point towards directions that can be more fully explored in verified short sleepers. This paper discusses problems associated with the terminology used to describe “short sleep,” summarizes laboratory studies exploring neurobehavioral performance, metabolism and obesity, and psychological health and epidemiological studies exploring mortality risk, obesity and metabolism, cardiovascular disease, and general health/psychosocial stress, describes studies of verified short sleepers and explores areas of convergence, laying out possible future directions.

Introduction

There is a substantial debate, both within the scientific community and society at large, about the minimum amount of sleep required for health and well-being. Central to this debate is a concern about the long-term effects of the shortening of sleep opportunity (sleep curtailment). There are no studies that directly address whether sleep curtailment is the same as sleep insufficiency. That is, is it the case that reduced sleep opportunity leads to reduced sleep time? While this seems likely, the relationship between changes in sleep opportunity and sleep time are not perfectly linear, as factors such as sleep ability (e.g., insomnia) and sleep homeostasis (e.g., sleep pressure) play a role. Additionally, if sleep time is reduced, is the reduction of sufficient magnitude to lead to negative outcomes? The evidence to suggest that sleep curtailment is associated with negative outcomes comes from two primary sources: (1) studies of acute sleep curtailment in the laboratory setting (sleep deprivation) ^1–18, and (2) epidemiological studies relating self-reported sleep to health outcomes ^{6, 9, 12–13, 19–27}.

The goals of his paper are to: (1) Review the existing literature examining “short sleep” using objective, laboratory-based methods, (2) Review the existing literature examining “short sleep” using epidemiological and other self-report methods, (3) Assess the degree to which there is a convergence between these approaches, and (4) Suggest further research required to clarify the underlying issue of whether short sleep is related to negative health outcomes.

A problem of terms: The lexicon of short sleep

An issue relevant to this discussion is the untangling of the various terms used to describe “short sleep.” In the popular press²⁸ and in the scientific literature²⁹ apparently synonymous terms (e.g., “short sleep,” “insufficient sleep,” “sleep loss”) are used interchangeably, although these terms have distinct, specific meanings that connote different methods and concepts. For the purposes of this paper, we have outlined working definitions for these and other terms in Table 1.

Table 1.

Working lexicon of Short Sleep terms

Term	Working Definition
Sleep Opportunity124:	Amount of time in which sleep is possible; including time in bed awake.
Sleep Time125Sleep Duration 126 –127:	Amount of time spent sleeping.
Insomnia128:	A clinical diagnosis with specific diagnostic criteria that describe a condition associated with pathologically long sleep latency, large amounts of wake time after sleep onset, and/or early morning awakenings associated with distress or impairment in functioning.
Partial Sleep Deprivation129:	Sleep deprivation in the laboratory setting, up to 5 hours sleep time; some sleep is achieved.
Self-Reported Short Sleep86	Subjective, retrospective estimate of habitual sleep time, or mean sleep time from prospective sleep measures (e.g., sleep diaries).
Short Sleep130:	Habitual sleep time of 6 hours or less.
Sleep Attainment131:	Habitual sleep quantity of sufficient quality.
Sleep Curtailment122:	The deliberate shortening of sleep opportunity.
Sleep Deprivation4:	Acute sleep curtailment in the laboratory setting.
Sleep Disturbance132–133:	Any problems associated with sleep, usually attributed to symptoms of diagnosable sleep disorders, such as insomnia, sleep apnea, narcolepsy, sleep-related movement disorders, and other sleep disorders; may apply to other clinical conditions, including affective disorder, anxiety disorder, or pain.
Sleep Restriction134:	Sleep deprivation (in or out of the laboratory), >5 hours sleep time but less than habitual sleep time. May in other contexts refer to a treatment for insomnia135.
Sleep Insufficiency136:	Reduction in sleep time of a magnitude associated with negative outcomes
Sleep Loss137:	Decrease in sleep time relative to previous sleep time
Total Sleep Deprivation138:	Complete elimination of sleep for 24 hours or more
Verified Short Sleep64:	Verification of self-reported habitual sleep duration using polysomnography or actigraphy

The question remains: What is “short sleep?” The phrase has been used to refer to all of the various terms in Table 1. It has been used to describe subjective and objective sleep duration, for both chosen and forced sleep opportunities, in both laboratory and home settings. In a way, it can mean all of those things, as, literally, the phrase is ambiguous. It doesn’t address how short the sleep is, what the shortness is relative to, and how it was determined. Thus, for scientific purposes, it is a suboptimal descriptor (necessitating the use of other Table 1 terms when they are appropriate).

However, we propose that there is a valid use of the term: To describe habitual sleep duration of <6 hours. Many of the epidemiological studies already use the term this way, and it helps describe subgroups of Short Sleepers in the population. For example, while it is possible that there are true Short Sleepers in the natural environment (those who demonstrate Short Sleep without Sleep Insufficiency), as well as those who demonstrate Short Sleep along with varying levels of Sleep Insufficiency, the existing literature has not specifically explored these groups. A clarification in terms may help build a nosological framework around which some of the heterogeneity among Short Sleepers can be explored.

To better understand what Short Sleepers (true or otherwise) may experience, we look to the existing Laboratory and Self-Report studies and examine their convergence, despite limitations.

Data from laboratory studies

One of the earliest laboratory studies investigating the role of sleep on performance was carried out almost a century ago by Pieron and colleagues³⁰, who were searching for biomarkers of sleep in dogs, using sleep deprivation. As the search for such biomarkers continues, the science of studying sleep has matured considerably, growing to explore many other hypotheses and research directions. Current studies employ laboratory-controlled sleep deprivation to examine outcomes associated with sleep loss, which may generalize to short sleep. Three domains in which this has been explored are: neurobehavioral performance, metabolism and obesity, and psychological health.

Laboratory Studies of Neurobehavioral Performance

Sleep deprivation leads to increased sleepiness. The term “sleepiness” is used to describe a state of increased sleep propensity (likelihood of falling asleep), usually referring to situations in which sleep is not appropriate or desired³¹. This is in contrast with “tiredness,” which is a subjective experience similar to fatigue. While they often overlap significantly, they are separate constructs – for example, individuals suffering from insomnia often experience tiredness, yet are not sleepy³². Individuals who experience partial or total sleep deprivation demonstrate increased sleepiness^{5, 33}. This finding has been replicated^{2, 5, 34–35} across three methods of objective assessment of sleepiness: the Multiple Sleep Latency Test, Maintenance of Wakefulness Test, and measurements of oculomotor activity⁵.

Within the paradigms of sleep deprivation and sleep loss, neurobehavioral performance is one of the most often-measured outcomes⁴, often operationalized as sustained attention. The dominant measure of sustained attention in paradigms of sleep deprivation is the Psychomotor Vigilance Task^{4, 36} (PVT). A consistent finding with the PVT is that sustained attention decreases with sleep restriction, and this decrease is exacerbated as sleep opportunity is reduced from 7 to 3 hours^4–5. Additionally, cumulative deficits across several days are seen when sleep schedules of 3, 4, 5, and 6 hours are maintained^4–5. Inferences, though highly important in our understanding of sleep duration and sustained attention, are limited to the subjects and conditions specified in these experiments. PVT performance, which measures one aspect of neuropsychological function, has not been evaluated outside of the laboratory. Other work has demonstrated that sleep deprivation is also associated with deficits in executive function³⁷, learning³⁸, and memory³⁹.

There has been significant discussion of the role of impaired functioning due to sleep loss in auto accidents^40–45. There is strong evidence that sleepiness is a major factor contributing to automobile accidents^46–48. Several papers have shown that accidents are related to sleepiness caused by sleep apnea^{46, 49–50} and prolonged wakefulness^{48, 51–52}. Some evidence also suggests that short sleepers may exhibit more sleepiness than average^53–54, but currently there is no data to support the claim that short sleepers are more likely to experience auto accidents⁵⁵.

Laboratory Studies of Metabolism

The first study to evaluate metabolic consequences of sleep loss was published over 15 years ago⁵⁶ in 8 healthy subjects. This study found that sleep deprivation had a negative impact on glucose tolerance and insulin resistance, but problems resolved with recovery sleep. Other classic studies found that rats who were continually deprived of sleep became hyperphagic, even though they demonstrated a depletion of energy reserves⁵⁷. More recent studies have expanded our knowledge about the relationship between sleep and metabolism.

Short sleep is associated with impaired glucose tolerance and insulin resistance. In a landmark study, Spiegel and colleagues¹⁴ studied the effects of sleep curtailment to 4 hours for 6 days, with baseline and recovery sleep opportunities. Compared to baseline, glucose tolerance decreased by 40%, glucose disposal decreased by 30%, acute insulin response decreased by 30%, and insulin sensitivity trended lower. Additionally, the area under the Homeostasis Model Assessment (HOMA) curve was 50% higher than when rested. This study had several limitations including small sample size, an order effect, and no control group. Nonetheless, it engendered significant interest in the field of sleep and metabolism and was followed by numerous studies. In a crossover design¹³ in which groups were allocated to receive 4 hours or 10 hours of sleep opportunity for 2 days each Spiegel and associates were able to replicate their earlier findings.

Other hormone systems are affected as well. Studies of sleep deprivation have also examined the hormones leptin (originates in adipose tissue and regulates appetite and energy metabolism) and ghrelin (released from th stomach and pancreas and stimulates appetite). These are hormones that, among other functions, regulate hunger and satiety^58–59. Total sleep deprivation has been shown to result in a decrease in the amplitude of the leptin rhythm⁶⁰, and a decrease in the normal rise in ghrelin associated with sleep⁶¹. In one study⁶², leptin decreased and its circadian phase was shifted earlier by 2 hours compared to baseline. A later study⁶³ replicated these findings in sleep deprivation and extension. Changes in leptin and ghrelin may be partially responsible for changes in hunger and appetite⁶³, as previous studies have shown that sleep deprivation is associated with cravings for more calorie-dense foods⁶³ and habitual shorter actigraphic sleep is associated with greater fat intake⁶⁴.

Laboratory Studies of Cardiovascular Health

While research in this area is relatively limited, there have been several sleep deprivation studies that have linked less sleep with poorer cardiovascular functioning. If there is a link between sleep deprivation and cardiovascular disease, it may be through CRP (described above) and other inflammatory markers⁶⁵. Coronary circulation may also be affected⁶⁶. Additionally, sleep deprivation results in blood pressure and sympathetic changes^67–68.

Laboratory Studies of Psychological Health

Sleep deprivation has effects on several domains of psychological health, including stress, mood and socialization. An increased stress response has been shown, including increased basal activity of neuroendocrine stress systems, elevations of the sympathetic nervous system, altered hypothalamic-pituitary-adrenal axis function, and increased stress reactivity¹⁶. Effects are modest, but chronic exposure may have gradual but cumulative effects. Exploring this issue further, Roman and colleagues⁶⁹ found that sleep deprivation may lead to changes in certain receptor systems that are either directly or indirectly related to stress response.

In a study by Haack and Mullington¹⁸, “optimism-sociability” declined 15% over consecutive days of sleep loss; additionally, a 3% (statistically significant) increase in bodily discomfort was reported, due to significant increases of generalized body pain, back pain, and stomach pain. These findings support other recent work showing a significant relationship between sleep deprivation and mood dysregulation¹⁷. These laboratory studies suggest that sleep loss is associated with increased depressive symptoms and stress. However, these findings need to be replicated in studies of verified short sleepers to ascertain whether they are the result of acute sleep deprivation or associated with habitual short sleep.

There is some recent evidence that induced sleep deprivation is associated with impaired decision-making. This may have to do with impaired executive functioning (described above). One study investigated risky decision making following sleep deprivation⁷⁰ and found that not only was sleep deprivation associated with increased risky behavior, but an effect was seen for age, such that older subjects were more likely to demonstrate a greater effect. Some studies have reported similar effects^71–73, while others have found mixed results⁷⁴.

Data from studies of self-reported short sleep

Epidemiological studies remain a valuable method for exploring relationships between sleep and health in the general population. These studies allow for the measurement of sleep and related health variables at a population level, elucidating relationships that may be too subtle to detect in laboratory studies but nonetheless valuable to society⁷⁵. The primary domains addressed by epidemiological studies that considered sleep are: mortality risk, obesity and metabolism, cardiovascular disease, and general health/psychosocial stress.

Epidemiological Studies of Mortality Risk

Over 40 years of epidemiological studies have examined the association between habitual sleep duration and risk of mortality^{19, 76}. These studies, spanning several decades and continents, and including millions of study participants, have replicated the pattern that “short sleep” and “long sleep” (with varied definitions across studies) are associated with increased mortality relative to those in the normative group – usually this consists of those sleeping 7–8 hours. In addition, studies comparing several sleep duration groups have generally found that the further the deviation from the normative range, the greater the increase in mortality risk¹⁹. A complete review of these studies is published elsewhere⁷⁷.

A recent meta-analysis⁷⁸ combined the results of these studies, and found that pooled relative risk (RR) for all-cause mortality for “short sleep” was 1.10 (95% CI 1.06, 1.15), and RR for all-cause mortality for “long sleep” was 1.23 (95% CI 1.16, 1.30). When cause-specific mortality was explored, the RR for cardiovascular-related was 1.06 (95% CI 0.94, 1.30) and cancer-related RR at 0.99 (95% CI 0.88, 1.13) in short sleep and cardiovascular RR was 1.38 (95% CI 1.13, 1.69) and cancer RR was 1.21 (95% CI 1.11, 1.32) for long sleep. This meta-analysis supports the claim that short sleep (which was defined variously in these studies but usually was defined as <7 or <6 hours) is associated with increased all-cause mortality risk, echoing the theme of many of the epidemiological studies. However, it should be noted that the RR for “long sleep” (usually defined as >8 or ≥9 hours in these studies) reflects a much greater associated risk than that for short sleep. Even if long sleep is associated with greater risk, the importance of the relationship between short sleep and mortality risk is not diminished. It has been asserted that the risks associated with long and short sleep represent two distinct phenomena which are to be considered separately^{77, 79}. Thus, these results demonstrate that short sleep is a predictor of mortality and that long sleep is also worthy of study regarding mortality risk.

Epidemiological Studies of Obesity and Metabolism

Many studies have shown a dose-response relationship of short sleep duration and obesity and metabolic consequences across all age groups^{9–10, 21, 25, 80}. Cross-sectional studies have been conducted in adults from Canada^23–24, France⁸¹, Germany⁸², Japan⁸³, Spain⁸⁴, the UK⁸⁵, and the USA^86–88, showing significant associations between short sleep and obesity.

Recent large studies have found several interesting relationships between self-reported short sleep duration and BMI. Increased BMI was associated with decreased sleep durations among men, but a U-shaped relationship was seen in women⁸⁶. Also, subjective time in bed was associated with BMI in a U-shaped manner after adjusting for age and sex, with the minimum BMI observed at 7.7 hours/night^87–88. Longitudinal analysis observed that sleep durations <7 hours were associated with increased risk of weight gain. In another study, data collected over 13 years showed that the odds ratio for sleep duration predicting obesity was 0.50, such that every extra hour increase of sleep duration was associated with a 50% reduction in risk of obesity⁸⁹. This finding has also extended from urban into rural samples⁹⁰. Studies outside the US have replicated these findings as well, in France⁸¹ and Canada^24–25.

Epidemiological Studies of Cardiovascular Health

Recent data suggests that individuals who reported less than 6 hours of sleep were at increased risk for hypertension^91–92, which found that those reporting habitual sleep of ≤5 hours were at greatest risk of incident hypertension; this relationship was attenuated in those reporting 6 hours. Other recent data ⁹³ was able to replicate this finding in women, but not in men. Additionally, women who sleep <7 hours were at increased risk of stroke⁹⁴, though the relationship for long sleep was stronger. Regarding myocardial infarction, short sleepers were at significantly increased risk⁹⁵. This relationship was notably stronger for women than men.

Studies of Subjective Short Sleep and General Health

Short sleep is associated with poor health in general. “Short” sleep (self-reported 6–7 hours or <6 hours) has been associated with lower reports of self-rated overall health than “normal” (7–8 hour) or “long” (8–10 or >10 hour) sleep in young adults²⁷. Short sleep has also been associated with low socioeconomic status^96–97, as well as living in an urban environment⁹⁷. These studies, however, did not examine whether these risk factors were associated with decreased health-related quality of life.

Also, short sleepers have been found to demonstrate shorter melatonin rhythms and more morning activity than long sleepers¹⁹. While implications of these effects are unclear, underlying circadian rhythm differences may suggest genetic influence. Additionally, self-reported short sleepers report higher rates of difficulty falling asleep, waking too early, waking during the night, waking unrefreshed, and feeling sleepy during the daytime than self-reported 7–8 hour sleepers⁹⁸. A study⁹⁹ found that short sleep was associated with drinking six cups or more of coffee per day, taking sleeping medications, difficulty initiating sleep, disrupted sleep, early morning awakening. Also, short sleep was associated with self-reported “Sleeping disorders or insomnia,” “difficulties in getting sleep without sleeping medicines,” “overexertion or exhaustion” and “being awake during the night.” It should also be noted that self-reported short sleepers usually includes people purposefully curtailing their sleep as well as individuals with insomnia who want to sleep more but are not able to do so. It may be the case that a large portion of self-described “short sleepers” are people with insomnia and/or are taking sleeping medications, which may play a role in health outcomes. Currently, there are no studies that have explored the heterogeneity among short sleepers to discern which are true short sleepers, which are experiencing insufficient sleep and which have sleep disorders.

Studies of Subjective Short Sleep and Psychological Health

Short sleep has been associated with increased anxiety and a coping style characterized as having more of an external locus of control than long sleepers¹⁰⁰. This means that short sleepers are more likely to look to external explanations for successes and failures. For example, they will be more likely to blame others for problems, and judge their successes based on the impressions of others. On the Ways of Coping Questionnaire, short sleepers reported higher “Confrontive,” “Self-controlling,” “Accepting responsibility,” and “Positive reappraisal” scores than long sleepers¹⁰¹. It is possible that this represents an underlying personality style that is linked to short sleep and related to the experience (or need) of less REM sleep¹⁰², but these effects have not been sufficiently explored in recent studies.

Self-reported short sleep has been associated with increased anxiety¹⁰³ and hypomania¹⁰⁴. In a study of reported stressful events, Sexton-Radek¹⁰⁵ found that short sleepers had more consistent responses to stress. This is consistent with an external locus of control. These studies suggest that preferred sleep schedules are associated with different coping styles. However, these studies did not verify whether the “short sleepers” studied did actually sleep less.

Data from verified short sleepers

Several studies have investigated health outcomes associated with sleep duration utilizing objective methods (polysomnography and actigraphy) to estimate sleep in subjects reporting habitual short sleep. While this literature is limited, it may clarify some of the findings from laboratory and epidemiological studies.

Sleep Homeostasis and Circadian Rhythms in Short Sleepers

In early studies of sleep duration^106–107, short sleepers were found to demonstrate more consolidated sleep, less REM (but increased REM density¹⁰⁸), stage 1 and stage 2 sleep, and equivalent or more slow wave sleep when compared to long sleepers.

There is some evidence that sleep homeostasis and circadian rhythms are fundamentally different in this group. Aeschbach and colleagues^{53–54, 109} explored these questions in several studies. First, they found that short sleepers and long sleepers show differences in recovery from sleep deprivation, suggesting a difference in sleep homeostasis⁵⁴. During baseline nights, short sleepers had a shorter sleep onset latency and higher sleep efficiency than the long sleepers. Additional differences appeared when the groups were compared after truncating all sleep recordings to the shortest sleep time, looking at the “longest common sleep interval.” When this is done, short sleepers still showed decreased sleep latency, but they also showed increased SWS, suggesting greater homeostatic sleep pressure in this group. Additionally, more REM sleep was seen in the short sleepers using this technique, suggesting chronic sleep restriction (resulting in REM rebound) or differences in circadian phase (resulting in earlier REM onset). Following sleep restriction, the long sleepers demonstrated a greater degree of impact on sleep. This landmark study was extended with later findings supporting the idea that short sleepers live under higher homeostatic sleep pressure, but they are much more adept at tolerating that pressure than long sleepers⁵³.

This group also examined whether short sleepers exhibit a shorter “biological night” than long sleepers. This study evaluated 10 long sleepers and 9 short sleepers in a paradigm of constant environmental conditions, without sleep, so that the expressions of endogenous biological rhythms could be assessed. This study specifically examined aspects of various circadian markers, including the melatonin, cortisol, and body temperature rhythms. Short sleepers demonstrated a significantly shorter duration of high-level melatonin secretion, cortisol secretion and low body temperature. Also, the well-established co-occurrence of the peak of cortisol secretion with the time of habitual awakening from sleep occurred 2.5 hours earlier in the short sleepers. The authors note that higher homeostatic pressure load may be an explanation of performance deficits seen in sleep deprivation, though it is unknown how this may be present in habitual short sleepers. These findings, as well as those by the earlier studies, suggest that the processes governing sleep homeostasis and circadian rhythms may be altered in habitual short sleepers.

Sleep Debt in Short Sleepers

Habitual short sleepers may be sleep deprived. In a recent study¹¹⁰ individuals recorded habitual sleep duration, then spent 3 nights in the laboratory, and were given as much sleep opportunity as they would take. This study examined 17 individuals of varying sleep lengths, and found that throughout the 3 nights in the laboratory, all of the participants slept longer than their habitual sleep patterns indicated, except for some of those with the longest sleep durations who, by the second or third night, were back to (or below) their habitual durations. This study included only 2 individuals with habitual sleep durations of 6 hours or less. When these individuals were given maximal sleep opportunity for three consecutive nights, they both slept for >10 hours all three nights – on the third night, these two subjects slept approximately 12 and 14 hours. This suggests that habitual sleep might harbor significant amounts of sleep debt, even in self-reported short sleepers¹¹¹. However, it should be noted that this was a very small study.

Coronary Artery Calcification

A recent analysis of the data from the CARDIA study¹¹², which obtained actigraphic measurements of habitual sleep, found that shorter sleep was associated with increased coronary artery calcification. In a linear regression analysis, each increase of 1 hour of sleep duration was associated with an OR of 0.67 (95% CI: 0.49–0.91). While a U-shaped relationship was not explored, these data suggest a possible mechanism through which sleep duration may be responsible for increased mortality risk.

Metabolic Disruption in Short Sleepers

The Wisconsin Sleep Cohort Study utilized sleep diaries, polysomnography and blood samples to examine relationships between sleep and health. Sleep time, when measured by polysomnography, was negatively associated with ghrelin levels (less sleep was associated with more ghrelin). Additionally, when sleep was measured by sleep diary, but not polysomnography, more sleep time was associated with increased levels of leptin¹¹³. All of these analyses adjusted for BMI. Thus, self-reports of less sleep was associated with decreased leptin (which may result in decreased satiety) and less polysomnographic sleep was associated with more ghrelin (which may result in increased hunger).

Diet of Short Sleepers

There has been one study of diet associated with both subjective and objective habitual sleep duration¹¹⁴. This study examined a subset of women enrolled in the Women’s Health Initiative (WHI), a large, multisite trial of health in postmenopausal women. Sleep and circadian rhythms were investigated in 459 women by collecting 1 week of sleep diaries and actigraphic recordings, and a sleep questionnaire (sleep recordings described previously¹¹⁵). Dietary data were based on the WHI Food Frequency Questionnaire¹¹⁶ that was completed as part of involvement with the WHI and analyzed by WHI nutritionists. Partial correlations adjusted for age, income, education, physical activity, BMI, and total gram amount of diet. Actigraphic total sleep time was significantly and negatively associated with total fat, monounsaturated fat, trans fat, saturated fat, polyunsaturated fat, total calories, gamma-tocopherol (a form of vitamin E, which is ingested almost exclusively in fats), cholesterol and alpha-tocopherol (also a form of vitamin E). There were no significant correlations with sleep diary total sleep time. It should be noted that the primary sources of vitamin E are the diet are oils. This suggests that there is a relationship between more fat intake (irrespective of type) and shorter habitual sleep duration. It should be noted that since this study is correlational, direction of causation cannot be measured. However, these findings support the hypothesis that metabolic obesity is closely related to sleep duration.

Convergence Between Experimental and Epidemiological Studies

Compared to the amount of data available from laboratory and epidemiological studies, there has been relatively little investigation of verified short sleepers. However, given the findings from these domains, several patterns emerge:

First, short sleep has been associated increased mortality risk. While the risk associated with long sleep may be greater, self-reported short sleep does carry an increased risk, which is unexplained by cancer and cardiovascular events, even though short sleep may be a risk factor for heart attack and stroke. The mortality relationship is a global issue, as studies from 5 decades and several countries have replicated this finding. Sleep may be directly related to mortality, or, more likely, it may mediate or moderate a relationship involving cardiovascular disease, obesity, metabolic dysregulation, stress, immune dysfunction, psychological health, cancer or coping difficulties. Laboratory results suggest that sleep deprivation is associated with impairments in these domains, and epidemiological studies confirm that short sleepers report impaired overall health as well as a number of cardiovascular and metabolic risk factors that support that this pattern may be seen in habitual short sleep. However, future studies need to link the laboratory findings and epidemiological results to verified short sleepers, so that the extent of the presence of these risks in this group is understood.

Second, short sleep has been associated with metabolic dysregulation and obesity. Sleep deprivation has been shown to produce short-term changes in a number of endocrine systems, including insulin, glucose, leptin and ghrelin. These differences may persist over the long term, explaining some of the epidemiological finding. Also, short sleepers may be more likely to eat higher-fat foods, which may be causing the obesity repeatedly demonstrated in epidemiological studies and may or may not be driven by endocrine changes.

Third, short sleep has been associated with worse cardiovascular health. While this has primarily been driven by epidemiological studies, with their inherent problems with the measurement of sleep (described below), the evidence suggests that those reporting less sleep are at greater risk of hypertension, stroke and myocardial infarction than those who sleep 7–8 hours. This relationship may be greater in women than men, and may be evident for long sleep as well. This is supported by a number of laboratory studies that suggest that sleep deprivation is associated with heightened blood pressure and sympathetic activity^67–68.

Fourth, short sleep has been associated with impaired neurobehavioral performance and cognitive functioning. These results have been primarily explored in the context of sleep deprivation studies and have not been replicated in naturalistic settings. It is unclear whether performance deficits associated with short-term sleep deprivation describe the experience of habitual short sleepers. These results are supported by epidemiological findings, which suggest that short sleepers report more sleep disturbance, including daytime sleepiness⁹⁸.

Fifth, short sleep has been associated with psychological/psychiatric disturbances and poor general health. Sleep deprivation studies show that at least short term neurophysiological changes indicative of stress and depressive symptoms result from sleep deprivation. Studies of self-reported short sleepers mirror these findings, showing that short sleepers exhibit more risk factors for stress and depression, as well as characteristic coping styles.

Problems with laboratory and epidemiological studies

Laboratory studies have several important limitations. First, these studies usually involve very small sample sizes. Many of these studies were of 10 individuals per group or less. Second, these studies do not study habitual short sleepers. Third, these studies are performed in an artificial setting. In reality, sleep is subject to socio-ecological influences at different levels⁷⁷. For example, caffeine, alcohol, smoking, children, pets or environmental noise may disturb sleep and some may not be present or allowed in a laboratory setting. Work and school schedules may play less of a role in the laboratory, and the experience of being with a typical bed partner and being exposed to his/her routines may influence sleep. Additionally, laboratory measures of sleep usually involve polysomnography, a method that requires equipment that may itself be difficult to adapt to for a night of sleep¹¹⁷. While this is most widely-known as the “first night effect,”¹¹⁸ the act of measuring sleep may disrupt sleep.

Epidemiological studies have also linked short sleep duration with several important health outcomes. However, there are a number of limitations that should be noted. First, inconsistency regarding the definition of “short sleep” reduces the generalizability of the conclusions across studies. Second, since these studies are primarily epidemiological, they capture self-reported measures of sleep duration, which may better reflect time in bed rather than actual sleep time¹¹⁹. Third, subjective and objective measures of sleep duration are often discrepant ¹²⁰, and both measures inherently contain error in their estimation. Fourth, these studies could not accurately describe how risks associated with sleep duration vary across the lifespan. While age was often used as a covariate, it is likely that the mechanisms by which sleep duration increases health risks have varying effects at different stages of life. Finally, these large studies fail to adequately measure the extent to which those reporting short sleep do so because of less biological need for sleep, rather than a deprivation of sleep¹⁹.

Future Direction: Investigate Short Sleepers

There is a large volume of research on the negative effects of shorter-term experimental sleep deprivation, and a growing public need for increased understanding of the role of sleep duration in several domains, including the prevention of disease ⁷⁵, driving accidents ¹²¹ and daytime functioning ⁶. This literature is extremely valuable as it has elucidated, in controlled environments, maladaptive processes that are seemingly caused by insufficient sleep in healthy subjects. However, it is difficult to extrapolate these findings to populations at large. In this regard, a dearth of literature exists to help us understand the effects of real-world shorter sleep durations.

Important differences exist between the laboratory-based studies and the epidemiologic and translational research needed to answer this important societal question. First, shorter sleep attainment in society is, to a large degree chosen, not imposed except in individuals with insomnia. Second, the exposure to shorter sleep may vary across days, months and years so the snapshot view obtained from most research studies may not be indicative of patterns over time. Third, and related, the effects of the exposure may occur and surface many months or years later. Thus, while the available experimental data are valuable for the generation of hypotheses regarding short sleep as it is experienced by members of the population, they do not sufficiently describe this group and research studying potential differences between short sleepers and sleep-deprived healthy adults needs to be undertaken.

There is a great need to investigate short sleepers, rather than “normal” sleepers who are deprived of sleep within a research setting – they may be at increased risk for a number of negative health outcomes, they comprise a large (and possibly growing) segment of the population, and little research has been devoted to this group. There are few known studies that have studied the sleep of habitual short sleepers and have investigated the performance, metabolic and biopsychosocial (e.g., mood, anxiety, functioning, behavior) characteristics of this group, relative to normative (7–8 hour) sleepers and relative to the sleep-deprived groups investigated in laboratory studies¹¹¹.

Considerations for future studies in short sleepers

Individual Differences in Habitual Short Sleepers

Not all short sleepers are the same. Recent data suggests that there are a number of reasons people follow a shorter sleep schedule; some believe that this schedule represents their natural way of functioning, and others trade sleep for time spent doing other activities; usually, sleep is traded for commute time and work¹²². Short sleepers are a heterogeneous group, comprising those that perceive impairment and those that do not, as well as those who believe that they are natural short sleepers, and those who choose this schedule to meet other demands. No previous studies have examined these characteristics of short sleepers, relative to objectively-estimated sleep. A future study might investigate who self-identifies as a short, normal and long sleeper, to validate the cutoffs suggested by the literature.

Some people are more sensitive to sleep loss. An important consideration is that sleep loss increases cognitive variability within subjects, such that certain tasks are more vulnerable than others, and significant differences have been reported between subjects ⁵. Sensitivity to sleep loss is reliable within subjects, but it depends on the task ¹²³. Thus, it is important to consider individual differences in response to sleep deprivation in the formulation of hypotheses regarding sleep loss. Regarding other neurocognitive processes, while there has been less investigation in these areas, there is evidence that sleep loss is associated with impaired memory and academic performance ³⁸.

Should Short Sleepers Simply Sleep Longer?

If habitual short sleepers demonstrate impairments, two important questions arise:

1. Does short sleep cause problems, or is it a result of them?

2. Does sleep extension alleviate these problems?

The first question is difficult to address: randomization to a short sleep lifestyle is not feasible, though other study designs may approach this question. The second question is the ideal next step for several reasons: it is more easily testable, would demonstrate causality in the other direction (supporting the claim that short sleep causes problems), and may be extended later to apply to acute sleep loss. Thus, with a sleep extension intervention, we could demonstrate if, in short sleepers, sleep extension improves health and functioning.

Practice Points

1. Terms describing various forms of “short sleep” need to be clarified and used appropriately. There has been much confusion among “sleep deprivation,” “sleep loss,” “insomnia,” and other related terms. This has led to confusion regarding interpretation of data. We propose definitions outlined in Table 1.

2. Laboratory studies have described impairments in neurobehavioral performance, metabolism and obesity, and psychological health associated with acute sleep deprivation.

3. Epidemiological studies have found links between self-reported short sleep duration and mortality risk, obesity and metabolism, cardiovascular disease, and general health/psychosocial stress.

4. Studies of verified short sleepers suggest that this group may experience greater homeostatic sleep pressure and sleep debt, exhibit cardiovascular risks, and are at increased risk for obesity.

5. These domains converge to suggest that habitual short sleep may be associated with increased mortality risk, performance deficits, cardiovascular risk, obesity risk and metabolic dysregulation.

Research Agenda

1. Findings from laboratory studies need to be replicated in verified short sleepers. This involves assessments of health, performance and psychological functioning.

2. Findings from epidemiological studies need to be replicated in verified short sleepers. Possible links between short sleep and mortality risks need to be explored in experiments rather than just epidemiological studies. Also, questions of how and why short sleepers may be at risk for obesity and diabetes need to be addressed in protocols involving verified short sleepers.

3. The “short sleepers” need to be phenotyped. This will be a heterogeneous group, and methods for differentiating short sleepers with decreased sleep need and those experiencing sleep insufficiency need to be explored.

4. Sleep extension interventions should be explored, once a valid approach for differentiating those with and without impairment is elucidated.