Short Sleep Duration Disrupts Glucose Metabolism: Can Exercise Turn Back the Clock?

Abstract

Short sleep duration is prevalent in modern society and may be contributing to type 2 diabetes prevalence. This review will explore the effects of sleep restriction on glycemic control, the mechanisms causing insulin resistance and whether exercise can offset changes in glycemic control. Chronic sleep restriction may also contribute to a decrease in physical activity leading to further health complications.

Summary

Sleep restriction causes insulin resistance, but the mechanisms are not well understood. Exercise may be an effective mitigator.

Introduction

Adults are advocated to receive at least 7 h of sleep each night to promote optimal health and reduce the risk of disease. In the last 50 years, the prevalence of inadequate sleep has almost doubled, with approximately 29–43% of the U.S. adult population currently failing to meet this recommendation (1,2). Even more concerning is that among high school aged students the prevalence is 70–84% (1). The factors contributing to the increased prevalence of sleep loss involves several broad societal changes, including longer working hours, shift work, alcohol consumption, lack of physical activity, and having greater access to television and the internet (3).

It has long been known that sufficient sleep is necessary for maintaining cognitive performance and work capacity, whereas insufficient sleep contributes to more traffic accidents, industrial accidents, medical errors, and loss of work productivity (4,5). It is estimated that on an annual basis, the U.S. loses an equivalent of 1.23 million working days due to insufficient sleep with an annual economic loss of $411 billion, equating to 2.28% of the gross domestic product (6). Besides impairing productivity, insufficient sleep duration has also been linked with seven of the fifteen leading causes of death in the U.S., including cardiovascular disease, malignant neoplasm, cerebrovascular disease, accidents, septicemia, hypertension, and type 2 diabetes mellitus (T2D) (7).

In recent years, particular attention has been focused on the metabolic implications that arise from chronically insufficient sleep. Despite the robust evidence highlighting that experimental sleep restriction impairs glucose tolerance by inducing insulin resistance (IR) in multiple tissues, little is known about the precise mechanisms attributable to this state. In particular the adipose tissue appears to play an important role in the IR. Additionally, the effectiveness of exercise to offset changes in glycemic control shows promise but has not been investigated thoroughly. We hypothesize that exercise may be a viable therapeutic intervention to mitigate sleep restriction induced IR (Fig.). Although research still needs to be conducted to establish how shortened sleep may alter physical activity patterns, we further hypothesize that chronic shortened sleep patterns contribute a decrease in physical activity exacerbating health complications.

Figure.

Schematic depicting overarching hypothesis. Insufficient sleep leads to metabolic dysfunction in skeletal muscle, adipose tissue, and liver, resulting in whole body insulin resistance. A lack of sleep may also disrupt several other physiological systems and promote a sedentary lifestyle, further exacerbating insulin resistance. When exercise is undertaken while sleep restricted, these outcomes may be avoided. Solid line = strong evidence. Dashed line = more research is needed. Created with BioRender.com

Sleep and glucose metabolism

Adequate sleep duration and quality are considered key lifestyle components along with diet and exercise in the management of T2D (8). The quantity of sleep is known to be associated (in a ‘U’ shaped manner) with health outcomes (e.g. HbA1c), with both long (>8 h) and short (<6 h) sleep durations having negative impacts (9). A recent meta-analysis of cohort studies revealed that sleeping ≤5 h/night and having poor sleep quality were associated with an increase in T2D risk by 45% and 38%, respectively (10). In 2017, the American Diabetes Association recommended for the first time that sleep pattern and duration assessments be part of the comprehensive medical evaluation of patients with T2D (11). A well-established predisposing factor in the development of T2D is IR (12) and several studies have implicated insufficient sleep as a risk factor for the development of IR (13–16). However, the causal relationship between sleep restriction and IR remains unclear.

In a seminal study, Spiegel and associates (17) studied the effects of sleep restriction (4 h/night for six days), baseline (8 h/night for three days) and recovery sleep (12 h/night) for 7 days in young healthy males. When administered an intravenous glucose tolerance test (IVGTT), the rate of glucose clearance during the sleep restriction condition was 40% lower when compared to the sleep recovery condition, with values similar to those observed in older adults with impaired glucose tolerance (18). Glucose effectiveness, the quantification of glucose’ ability to mediate its own disposal independently of insulin, was 30% lower during sleep restriction compared to the sleep recovery condition. The difference was comparable to that reported between groups of patients with T2D and individuals with normoglycemia (19,20). During the sleep restriction condition, the acute insulin response was 30% lower as compared to the well-rested conditions, with the difference in magnitude comparable to that described in ageing and gestational diabetes (21,22). Differences in insulin sensitivity were not significant between conditions but tended to be lower when sleep was restricted. Although there were several limitations including a small sample size, and failure to include a control group or to counterbalance order of assessment, Spiegel and colleagues were able to replicate their earlier findings in a crossover design study, where groups received 4 h or 10 h of sleep opportunity for two days (14,17).

The work of Speigel provoked significant interest in the field of sleep restriction and metabolism and was followed by several studies with both similar and conflicting results (23–29). Previously, other studies involving sleep loss showed that periods of total sleep deprivation ranging from 60–120 h promoted a prediabetic phenotype in otherwise healthy volunteers (30,31). The relatively severe and short-term sleep restriction raised questions about the generalizability of the findings. Subsequently, Zielinski et al., (23) studied older self-reported long sleepers and restricted sleep to 1 h less/night for eight weeks, and found no significant changes in fasting glucose, 2 h glucose following an oral glucose tolerance test (OGTT), or insulin sensitivity when sleep was restricted. Methodological differences likely attributed to the discrepancies observed. First, previous work (17) assessed young adults with average sleep durations whereas Zielinski assessed older self-reported long sleepers. Second, studies have employed profound sleep restriction protocols or total sleep deprivation for no longer than six consecutive nights (17,30,31) while the protocol utilized in the Zielinski study was relatively modest over a prolonged period of time. The discrepant results may reflect dose-response or threshold effects of sleep restriction on glucose/insulin regulation, or they could potentially reflect an adaptation that occurs with more chronic sleep restriction.

To investigate whether an adaptation occurs, Roberston et al., (32) assessed the time-course of the metabolic changes in response to moderate sleep restriction (1.5 h/night less than habitual sleep duration for three weeks). Weekly assessments (via clamp) revealed an initial decrease in insulin sensitivity after one week which then recovered to baseline levels by week three. More recently, Sweeney (33) investigated whether there is a cumulative effect from multiple nights of sleep restriction on glucose tolerance. Healthy participants underwent four nights of control (8 h) and restricted sleep (4 h) and were administered an OGTT each morning. There were higher insulin concentrations with sleep restriction but there was no evidence that the impairment in glycemic control increased in a cumulative manner with subsequent sleep-restricted nights. Although the intervention was short in duration, the findings suggest that the impairment in insulin sensitivity occurs at the onset of sleep restriction and may show a compensatory response over time.

In another study, Nedeltcheva et al., (24) examined whether recurrent bedtime restriction would result in decreased glucose tolerance and reduced insulin secretion and action. Healthy middle-aged males and females underwent two 14-day periods of controlled exposure to sedentary living with ad libitum food intake and 5.5 or 8.5 h time in bed. Following sleep restriction, glucose tolerance (via OGTT) decreased by ~10% and insulin sensitivity (via IVGTT) decreased 17%. Interestingly, glucose effectiveness increased by 15%, which in association with reduced insulin sensitivity has been reported in healthy offspring of patients with T2D and is thought to represent an important compensatory mechanism for the maintenance of normal glucose tolerance (34).

Of note, the majority of sleep restriction studies primarily use male participants and consistently report significant findings (17,26,35–39). In trials comprising both male and female participants, there is no unanimity as some report that sleep restriction causes glucose dysregulation (24,25,28,33), whereas others find no such relationship (23,29,40). The conflicting results suggest that sex differences may exist and that females could be somewhat protected. There is some evidence that females are more insulin sensitive than males matched for age and BMI (41,42). It is suggested that this protection is due to females having greater adipose tissue (43) and hepatic (44) insulin sensitivity. Indeed, a higher basal as well as insulin-stimulated methyl-glucose uptake has been observed in vitro in female versus male subcutaneous adipocytes (45). Differences in adipose tissue distribution likely favors females as gluteal adipocytes are more insulin sensitive with respect to inhibition of lipolysis than abdominal subcutaneous adipocytes (46). To date, no study has been sufficiently powered to assess sex differences in insulin sensitivity following sleep restriction. One of the few reports involving both sexes (eight males/six females) (27) observed a trend toward a ~30% greater reduction in whole body insulin sensitivity in males, but did not reach statistical significance, likely due to a small sample size. Furthermore, the female participants typically studied are premenopausal, but a recent report indicated that the possible sexual dimorphism is ameliorated post menopause (47–49). Singh et al., (47) noted reduced insulin sensitivity (via clamp) in postmenopausal women following four nights of sleep restriction (5 h /night). Bosy-Westphal et al, (48) reported no change in glucose tolerance (via OGTT) in pre-menopausal women following four nights of consecutively increasing sleep restriction. Together, the findings suggest that postmenopausal women may be more affected by sleep restriction than their premenopausal counterparts. Recently Zuraiki et al., (49) compared the causal effects of sleep restriction (1.5 h less than habitual sleep duration for six weeks) in pre- and postmenopausal women. In the full sample, sleep restriction did not significantly affect any of the OGTT-derived measures of glucose metabolism. However, when participants were stratified by menopausal status, calculated whole-body insulin sensitivity (via Matsuda index) tended to decrease in sleep restricted postmenopausal, but not premenopausal, women. Menopausal status also influenced the effect of sleep restriction on HOMA-IR which increased to a greater extent in postmenopausal, relative to premenopausal, women.

Another noteworthy point is that findings in this area vary depending on the method used to assess glucose metabolism. Each technique has its strengths and their limitations and have been described previously (50). No effect of sleep restriction has been reported when participants were assessed using an OGTT (23,29), fasting blood samples (29,40), or hyperinsulinemic euglycemic clamp (13). In contrast, others have shown an impairment in glucose metabolism in response to sleep restriction when utilizing an OGTT (24), IVGTT (17,35), or continuous glucose monitoring (CGM) (36). The conflicting findings between studies may be due to the severity of sleep loss during the imposed interventions. Further differences could be accounted for by the duration of sleep allowed during the control sleep period which can be a set duration or habitual sleep durations of the participants.

Taken together, the widespread lack of sleep in modern society is likely contributing to the prevalence of T2D, as IR consistently presents under conditions of experimental sleep restriction. As studies have predominantly used males as participants, there is a great need for future studies to be powered sufficiently to detect potential sex differences and to identify vulnerable populations.

Tissue specific IR

Although studies show a decrease in glucose tolerance or increased insulin concentrations with sleep restriction, the tissue specific mechanisms are unknown. Insulin resistance can be characterized by decreased insulin mediated glucose uptake in peripheral tissue, while in the liver IR could be due to augmented endogenous glucose production (EGP). In adipose tissue, a blunted suppression of lipolysis, increased non-esterified fatty acid levels (NEFA) (51), and altered fatty acid oxidation (52) are associated with IR. With conditions such as obesity (53–55) or behaviors such as physical inactivity (56–58), IR manifests in each of these tissues. To characterize the effects of sleep restriction on insulin sensitivity in specific tissues, Donga et al., (25) utilized hyperinsulinemic euglycemic clamp conditions with isotope dilution of a glucose tracer in healthy subjects following a single night of sleep restriction and control sleep (4 h and 8.5 h, respectively). Compared with control sleep, sleep restriction did not alter basal levels of glucose, NEFA, insulin, or EGP, however, during the clamp conditions, sleep restriction resulted in a 22% increase in EGP, indicating hepatic IR and a decreased rate of glucose disposal (~20%), reflecting impaired peripheral insulin sensitivity. Plasma NEFA levels were increased by ~19%, demonstrating decreased insulin sensitivity of lipolysis. Lastly, the rate of infusion of glucose during the clamp necessary to maintain glucose levels was ~25% less when subjects were sleep restricted.

Using a similar protocol, Rao and colleagues (27), examined insulin sensitivity to multiple nights of sleep restriction. Subjects had either 8 h of sleep or 4 h of sleep for five consecutive nights. Fasting glucose and insulin levels were similar following both conditions. During an OGTT, glucose levels, glucose AUC, and insulin levels were also not significantly different between conditions. During hyperinsulinemia (via clamp), whole-body and peripheral insulin sensitivity were reduced by 25% and 29% respectively, with sleep restriction. In the liver, EGP during hyperinsulinemia was not worsened by sleep restriction. Notably, circulating levels of NEFAs were 62% higher during sleep restriction and were accompanied by an increase in fat oxidation (via indirect calorimetry), a 24% decrease in triglycerides, and a non-statistically significant 55% increase in β-OH butyrate, a measure of hepatic fat oxidation.

Recently, our group (29) investigated the impact that mild sleep restriction (6 h/night for five nights) had on glucose and lipid metabolism in adults with obesity. Sleep restriction increased peak glucose and insulin concentrations but had no effect on glucose AUC in response to a mixed-meal tolerance test. In accordance with previous work (27), fasting NEFA concentrations were elevated following sleep restriction which coincided with enhanced fat oxidation indicating a whole-body switch toward fatty acid metabolism. Abdominal adipose tissue samples were taken to explore differences in lipolytic capacity and showed that fatty acid synthase protein content tended to increase with sleep restriction. This may contribute to the development of IR as excessive fatty acid accumulation is known to disrupt insulin signaling (59). Together, these studies demonstrate that the negative ramifications of sleep restriction are profound and are evident in multiple tissues.

Molecular mechanisms

Although often underappreciated in metabolism, adipose tissue is an important player in the regulation of glucose metabolism (60–62). Impaired adipocyte function is recognized as one of the main mechanisms linking sleep restriction with abnormal metabolic effects. The first study to report a mechanistic path was by Broussard et al., (63) who collected adipose tissue biopsies and performed ex vivo studies on adipocytes following sleep restriction in humans. They demonstrated that four days of sleep restriction (4.5 h total sleep) decreased the half-maximal pAkt/tAkt response to insulin and the total AUC of the pAkt/tAkt response was 30% lower than during normal sleep indicating that sleep restriction induced adipocyte IR. With the reduction in adipose tissue insulin signaling was a concomitant ~16% decrease in whole body insulin sensitivity (via frequently sampled IVGTTs). Although no study since has reported a mechanistic path in adipose tissue, Wilms et al., (64) noted that acute sleep restriction (4 h for one night) induced profound restructuring of the morning to evening adipose tissue transcriptome, resulting in disturbed glucose homeostasis, as characterized by reduced β-cell capacity.

As skeletal muscle accounts for up to 80% of glucose disposal under insulin-stimulated conditions (65), it has been the tissue of interest in mechanism-driven sleep restriction studies (37,66,67). Sweeney et al., (37) investigated whether changes in whole-body insulin sensitivity following acute sleep restriction (4 h/night for two nights) were accompanied by changes in Akt activity in skeletal muscle when compared to habitual sleep duration. During an OGTT, there was an increased plasma insulin response following sleep restriction, but the plasma glucose response did not significantly differ between conditions. No difference in skeletal muscle Akt activity was found between conditions in response to the OGTT, suggesting that only two nights of shortened sleep may not impact skeletal muscle Akt or insulin signaling. These results vary from Broussard (63) who reported changes in Akt activity when exposing adipocytes to insulin in vitro. While an OGTT demonstrates ecological validity, skeletal muscle exposure to circulating insulin can vary between participants. In another study, Saner et al., (38), investigated mitochondrial content and respiratory function and observed reductions in both with sleep restriction (4 h/night for five nights). During the OGTT, there was an increase in total glucose AUC and mean insulin AUC (22% and 29% respectively) following sleep restriction as compared to the control group (8 h/night for five nights). There was no significant change in insulin AUC with either intervention, despite a 29% increase in mean insulin AUC following sleep restriction. No differences in GLUT4 protein content or mRNA were observed between groups (38) nor were there changes in Akt activity reported. They observed a reduction in maximal mitochondrial respiration and significantly lower sarcoplasmic protein synthesis (a proxy for mitochondrial protein synthesis) in the sleep restricted group only, while no change in mitochondrial content was observed in either group. Although the parallel group design is a major limitation of this study, the authors do provide evidence that there is a concomitant decrease in mitochondrial respiratory function/synthesis and glucose tolerance following sleep restriction in otherwise healthy subjects. It has also been shown that sleep restriction (66) and sleep deprivation (67) promote a catabolic environment, decreasing myofibrillar protein synthesis rates (19% and 18%, respectively), which over time would likely accelerate muscle loss. Indeed, while in an energy deficit, curtailing sleep opportunities to 5.5 h/night for 14 nights increased the fraction of muscle mass lost compared to participants who had an 8.5 h sleep opportunity each night (68). Sleep restriction itself did not decrease muscle mass but did augment the loss during an energy deficit. A similar study (69), utilizing a less severe, more prolonged, model of sleep restriction supported this finding as individuals lost more muscle mass and retained more adipose tissue when sleep restriction occurred during an energy deficit. As greater muscle mass allows for increased skeletal muscle glucose uptake and improved insulin sensitivity (70–72), loss of muscle mass exacerbated by sleep restriction may contribute to IR and T2D development.

It is clear that both adipose tissue and skeletal muscle lose their sensitivity to insulin with sleep restriction but appear to do so at independent points along the insulin signaling pathway (e.g., Akt activity in adipose tissue but not in muscle). Although the point of dysregulation has not be delineated in muscle, several other metabolic pathways are impeded in muscle which may contribute to the resulting IR.

Other mechanisms

Other indirect mechanisms have been implicated with the negative impact of sleep restriction on health. The impact of sleep restriction on appetite regulating hormones such as ghrelin (68,73), leptin (73–75), glucagon-like peptide 1 (40), peptide tyrosine tyrosine (76), and regulation of food intake and energy expenditure (77–80) likely affect glucose metabolism indirectly by promoting weight gain and subsequent IR. Also a decrease in brain glucose utilization, concomitant with the drop in cognitive performance, that occurs with total sleep deprivation could potentially contribute to the disruption of whole-body glucose utilization (81). Researchers have also consistently shown that sleep restriction increases sympathetic nervous system activity (17,24,27,82,83) which is known to regulate insulin and glucagon secretion (84).

Sleep loss is a profound physiologically relevant stress on the body and thus stimulates the hypothalamus-pituitary-adrenal (HPA) axis. Sleep loss effects on hormones released from the HPA that regulate blood glucose is a critical link to increased risk of IR. Cortisol release induces gluconeogenesis, reduces peripheral glucose utilization, induces IR, and raises blood glucose concentrations (85). Several studies have shown that sleep restriction significantly increases salivary and serum cortisol levels during the late afternoon to early evening hours (17,40,86); these are times when cortisol is on the descending limb of its circadian rhythm, whereas other studies showed little to no changes (24,26,87). However, increased cortisol levels do not necessarily correlate with changes in insulin sensitivity (40). Following sleep restriction, prolonged nocturnal secretion of growth hormone (GH), a well-established counterregulatory hormone against insulin action (88), could promote morning IR. Broussard et al., (39) showed that sleep restriction for four nights (4.5 h/night) relative to normal sleep (8.5 h/night) increased NEFA levels during the nocturnal and early-morning hours which was related to extended nocturnal secretion of GH and a reduction in estimated insulin sensitivity (via IVGTT).

Short term sleep loss affects inflammatory homeostasis and chronic sleep restriction leads to a systemic increase in the concentration of inflammatory mediators that may have prognostic significance for metabolic disease (89,90). Acute sleep restriction (24), or sleep restriction imposed along with opportunity for intervening daytime naps (91), is not sufficient to increase subsequent daytime circulating levels of inflammatory markers. Despite this, one night of sleep deprivation has been found to induce increases in monocyte stimulated interleukin-6 (IL-6) and tumor necrosis factor α (TNFα) mRNA (92). Both IL-6 and TNFα are studied in this area due to their association with IR. IL-6 is a proinflammatory cytokine that causes IR by impairing the phosphorylation of the insulin receptor and insulin receptor substrate-1 (93). Similarly, TNFα decreases insulin-dependent glucose uptake by inhibiting autophosphorylation of the insulin receptor (94). Acute sleep deprivation (40 h) is sufficient to observe elevations in circulating levels of TNFα (95), whereas prolonged sleep deprivation (88 h) is needed to increase IL-6 (96). For sleep restriction, the length of the intervention must be longer in duration (7–12 nights) for a significant increase in circulating levels of inflammation to occur (97–99). However, findings are mixed as a recent meta-analysis found that sleep restriction had no overall effects on circulating levels of IL-6 and TNFα (100). Nonetheless, when insulin sensitivity is measured alongside markers of inflammation in these studies, an association between the two appears to exist (101). As increases in circulating levels of inflammatory cytokines can arise from nonimmune sources such as adipose tissue, it has been a tissue of interest in sleep restriction research.

Sleep restriction directly causes systemic IR but also disrupts several other physiological processes which could amalgamate and augment the IR response.

Exercise as a mitigator

While a single day and even modest sleep restriction negatively influences insulin sensitivity, little work has focused on minimizing these negative consequences. The broad range of health benefits that are attributed to regular physical activity/exercise are well documented and understood (102,103). Exercise is a particularly potent tool used for the prevention and in the treatment of T2D (104,105) due to its insulin sensitizing effects across tissues (106–108) and its ability to stimulate insulin-independent glucose uptake in skeletal muscle (109). Several of the aforementioned mechanisms that are potentially disrupted with sleep restriction, particularly insulin signaling, share similar biochemical signaling pathways to those connected with exercise. Because of this, we hypothesize that exercise may be a viable therapeutic intervention to mitigate sleep restriction induced IR.

Few studies to date have investigated the metabolic effects of sleep restriction with or without exercise (30,38,66,110–113). VanHelder et al., (30) were the first to report on the metabolic impact exercise has during a period of total sleep deprivation. Using a crossover design, healthy males underwent 60 h of sleep deprivation while remaining sedentary throughout or engaging in daily exercise. No differences in the total plasma glucose response to an OGTT were observed between conditions but the insulin response was elevated following both conditions. Interestingly, the sedentary condition resulted in higher insulin responses at all times compared to the exercise condition. It was already well established that sleep deprivation would induce IR but for the first time this study demonstrated that it could be partially reversed by concomitant exercise.

Subsequently de Souza et al., (110) investigated the effect of two weeks of high-intensity interval training (HIIT) prior to sleep deprivation (24 h). The training consisted of six sessions containing 8–12 repeated efforts for high intensity cycling. Young healthy males were submitted to four different conditions: a single night of regular sleep (8 h), sleep deprivation, HIIT followed by regular sleep, and HIIT followed by sleep deprivation. Total sleep deprivation increased both glucose and insulin AUCs during the OGTT, however, HIIT before sleep deprivation attenuated these increases. Prior HIIT was also able to partially restore NEFA levels that were elevated following sleep deprivation, potentially due to mitochondrial training adaptations that allowed for skeletal muscle to better utilize fatty acids for energy (114).

While exercise prior to and during periods of sleep loss were shown to be have protective effects, Sweeney et al., (111) investigated the effect of a single bout of sprint interval exercise on whole-body glucose metabolism and insulin sensitivity, following a single night of sleep restriction (4 h). The sprint interval exercise consisted of four all-out 30-s sprints against 7.5% of body mass, interspersed with 4.5 min of recovery which was previously shown to increase insulin sensitivity during an IVGTT by 142% (115). Healthy males underwent four conditions mirroring de Souza’s study design. Using an OGTT, there was no effect of condition on total glucose or insulin AUCs; however, late-phase insulin AUC was significantly lower with sleep restriction + exercise compared to sleep restriction alone. As the late phase of an OGTT can predict incident of T2D independent of the early phase (116), exercise after a single night of sleep restriction may be an effective mitigator.

Likewise, Saner et al., published several reports examining the effect of HIIT during consecutive nights of sleep restriction (38,66,112). Studying the same cohort, healthy males were subjected to sleep restriction (4 h/night for five nights) with or without HIIT (3 sessions consisting of 10 × 60s intervals at 90% W_peak). One study reported (66) that skeletal muscle myofibrillar protein synthesis was decreased by 19% following sleep restriction and was completely mitigated with exercise. The second paper (38) demonstrated that sleep restriction impaired glucose tolerance (via OGTT) and attenuated skeletal muscle mitochondrial respiratory function and synthesis, but the perturbations were not observed with exercise. They also noted that sleep restriction increased enrichment of inflammatory-related transcriptional pathways and decreased pathways associated with mitochondrial regulation in skeletal muscle while these changes were absent with exercise (112). Together, these findings suggest that the impairments caused by sleep restriction can be eradicated with high intensity exercise, particularly in skeletal muscle.

Light physical activity and moderate-intensity exercise appear to be insufficient in mitigating sleep restriction-associated metabolic impairments. Vincent et al., (113) reported on the impact of breaking up prolonged sitting with light-intensity walking breaks on measures of glucose metabolism (via CGM). Healthy males were sleep restricted (5 h/night for three nights) while remaining sedentary or completing a 3-min bout of light intensity walking every 30 minutes across a 7 h period. Irrespective of experimental condition, interstitial glucose was elevated following the first night of sleep restriction and did not differ between conditions throughout intervention. Thus sleep restriction seems to offset the insulin sensitizing effect of breaking up prolonged sitting with light physical activity (117) and that exercise must be performed at a higher intensity to counteract the insult sleep restriction imposes. Porter et al., (29) compared the effect of modest sleep restriction (6 h/night for five nights) with or without daily moderate intensity exercise (45 min walking at 65% VO_2max) on metabolic parameters. Exercise effectively decreased the insulin response to a meal but did not prevent the increased peak glucose response or elevated NEFA levels caused by sleep restriction. Together, these findings suggest that exercise must be performed at higher intensities to mitigate the robust metabolic dysregulation that sleep restriction elicits.

The mechanisms that account for this difference in efficacy are not known. Meta-analyses indicate that both HIIT and moderate intensity continuous training (MICT) are effective for reducing postprandial glycemia and insulinemia, particularly in those with impaired glucose metabolism (118), however, a more beneficial effect of HIIT has been observed on insulin sensitivity (119). Both changes in intramyocellular lipids and lipid intermediates (120) and glycogen depletion (121) after acute exercise have been linked to improved insulin sensitivity. It could be inferred that HIIT leads to more robust changes in these than MICT, therefore improving insulin sensitivity to a greater extent but that is not the case (122). However, a comparison of both training styles under sleep restriction has not been made and warrants further investigation.

Based on limited research, HIIT appears to be the most potent mitigator of IR in the setting of sleep restriction. Other intensities, durations, and modes of exercise should not be dismissed as more trials evaluating exercise effectiveness on sleep restriction induced IR still need to be conducted.

Does sleep restriction affect physical activity?

Exercise appears to be an effective tool at attenuating the negative consequences of sleep restriction; however, it is important to acknowledge that individuals may not be able to engage in exercise at the optimal intensity required, or that they may decrease their physical activity, when sleep restricted. A lack of physical activity impacts several metabolic tissues negatively (123), and these effects are likely compounded when physical inactivity is accompanied by sleep restriction.

The influence of sleep restriction on exercise performance and its adaptations has received considerable scientific attention. For those who are physically active, sleep restriction may influence acute metabolic training adaptations due to impaired performance (124). Impaired performance likely results from a decrease in muscular strength (125), endurance (126) and an increase in perceived effort (127). Overall, sleep restriction negatively impacts next day exercise performance, but the magnitude of the impact depends on the type of exercise performed and the manner in which sleep was restricted (128). Total sleep deprivation and restricting sleep through early awakening have a larger negative effect on exercise performance than restricting sleep by delaying bedtime. Further, exercise performed in the evening is more likely to be affected by sleep restriction than exercise performed in the morning. Also, the length of time awake prior to exercise is an influential factor. Taken together, several factors should be considered, but whether individuals will engage in exercise when sleep restricted is the most important thing to consider.

It has been hypothesized that chronic sleep restriction reduces energy expenditure, leading to weight gain (129). Despite this, 24 h energy expenditure (via whole room calorimetry) does increase when sleep restricted, but this is predominantly driven by the energy cost of extended wakefulness as opposed to physical activity (130,131). Given that more time spent awake accompanies sleep restriction, it could be argued that individuals can either use this time to increase their energy expenditure through physical activity or refrain from it and thus are inactive due to fatigue from a lack of sleep.

Evidence surrounding these hypotheses are conflicting as some researchers have found that physical activity decreases (132–134), increases (77,135,136), or does not change (137–139) when sleep is restricted. As seen with measures of glucose metabolism, disparities may arise due to methodological differences such as the research setting (e.g., free-living vs laboratory conditions). It is expected that participants will be more physically active under free-living conditions compared to the confines of a sleep laboratory. Therefore, the likelihood of capturing significant change in physical activity in a lab setting is low as opportunities for physical activity are low in the first place. Calvin et al., (138) reported no difference in activity counts (via accelerometry) between individuals during sleep restriction (5 h/night) and control sleep (7.5 h/night) for eight nights in a laboratory setting. Covassin et al., (139) also reported no differences between participants undergoing sleep restriction (4 h/night) and control sleep (9 h/night) for 14 nights in the same setting. In contrast, healthy adults with a familial risk factor for diabetes, had 31% fewer total activity counts and spent 24% less time engaged in moderate plus vigorous physical activity when sleep-restricted (5.5 h/night for seven nights), while in a laboratory setting where indoor leisure activity was allowed (133). In the only study that has compared these parameters in both a free-living and laboratory setting, Schmid et al., (132) reported that healthy adults reduced free-living, spontaneous physical activity after one night sleep restriction (4 h) compared to control sleep (8 h), specifically with a shift to lower physical activity intensities. However, for the second day/night of sleep restriction, participants stayed in a research unit and there was no change in activity between both conditions. Under both sleep conditions, physical activity was ~50% lower in the research unit compared to free-living, highlighting that the setting can influence behavior significantly.

Under free-living conditions, reports on physical activity during or following a period of sleep restriction have mixed findings. The manner in which sleep is manipulated (e.g., sleep timing and duration) likely influences these results. Brondel et al., (135) were the first to report increased free-living physical activity of healthy men that were sleep-restricted (4 h/night for 1 night). Specifically, physical activity during the evening hours was elevated compared to one night with 8 h of sleep, despite increased perceived sleepiness. In contrast, Tajiri et al., (137) reported increased sedentary time in healthy women when sleep was restricted (4 h/night for three nights), although this significant effect was clearly attenuated after adjustment for awake time. Due to contradictory findings, researchers have investigated if activity behaviors are affected by the timing of sleep loss.

In the first report of this kind (136), it was found that moderate-intensity physical activity time was greater when bedtime was delayed compared to habitual sleep times and advanced wake-time. In contrast, vigorous-intensity physical activity time was greater following advanced wake-time vs. delayed bedtime. Wilms et al., (134) added to these findings as they showed that during advanced wake-time sleep restriction, physical activity was substantially decreased compared to regular sleep and delayed sleep loss. Most recently, we showed that daily steps tended to increase following advanced wake-time compared to habitual sleep, while steps during the delayed sleep condition were unchanged, with no differences between sleep restriction conditions (77). Together, these reports show that sleep restriction can impact subsequent physical activity engagement in a free-living setting but also highlight that how sleep is curtailed influences those behaviors.

Together, these reports show that sleep restriction can impact subsequent physical activity engagement in a free-living setting, but also highlight that how sleep is curtailed influences those behaviors.

Future Perspectives

As discussed in the paragraphs above, there is strong evidence to suggest that sleep restriction affects insulin sensitivity in multiple tissues. Given that the evidence for molecular mechanisms is still preliminary, additional investigations utilizing tissue biopsies and gold standards methods for assessing insulin sensitivity (i.e., clamps) are required for robust conclusions. Further, research needs to be done using populations other than young healthy males (e.g., older adults, individuals with obesity, females) to delineate whether certain groups are more at risk than others. Future studies should also consider the manner in which sleep is restricted (e.g., early wake time or delayed bedtime) as it has been shown to determine subsequent behavior in the following day (77) and can impact glucose homeostasis (140). Additionally, exercise has shown promise as a tool to mitigate the health implications that result from sleep restriction; future studies should incorporate exercise as an additional arm (e.g., normal sleep, sleep restriction, sleep restriction + exercise) to elucidate whether it has efficacy under different sleep conditions. Consideration should also be given for when exercise is undertaken. Several studies have reported that evening exercise is superior to morning exercise for improving fasting and nocturnal glucose and postprandial insulin (141), reducing CGM-based glucose concentrations (142), and decreasing postprandial glucose (143). Comparing the impact of exercise timing during sleep restriction has never been investigated. Insight into whether evening exercise should be recommended over morning exercise for sleep deprived populations is a question worth researching.

Countermeasures to sleep restriction other than exercise should also be evaluated for their effectiveness in combating IR. An anecdotal solution for a lack of nocturnal sleep is daytime napping but given methodological difficulties with implementing, controlling, and quantifying naps, they have yet to be investigated in this experimental context. Cross-sectional and prospective cohort studies have associated daytime napping with dysregulated glucose metabolism when nocturnal sleep duration is adequate but beneficial when sleep is curtailed (144,145). An emerging solution, light therapy, which involves individuals being exposed to bright light early in the morning, is traditionally assumed to act by entraining the sleep-wake cycle via ocular stimulation of the brain’s suprachiasmatic nucleus (SCN), thereby improving sleep and circadian rhythmicity (146). As the SCN plays a main role in regulating glucose metabolism, it has been hypothesized light therapy may directly or indirectly improve insulin sensitivity in those with curtailed sleep. Although light therapy has been shown to improve sleep architecture and circadian rhythmicity, no discernable improvements in insulin sensitivity or blood glucose levels have been reported (147,148). However, its effectiveness during experimental sleep restriction has never been assessed, therefore its therapeutic value in that context cannot be ruled out. Another parameter that should be accounted for in future studies of this nature is dietary intake. Sleep restriction causes individuals to increase their daily caloric intake and reduce adherence to healthy eating patterns (73). Although a dietary intervention has yet to be investigated as a countermeasure, caloric restriction can result in increased insulin sensitivity in individuals with obesity and potentially contribute to regulating blood glucose levels and mitigating IR (149). Whether caloric restriction can mitigate the onset of IR during sleep restriction is not known. We believe the potency of exercise as a mitigator should not be overlooked and should be the primary focus in future studies as even low-intensity physical activity can significantly improve postprandial glycemic response, independent of energy balance (150).

Key points.

• Habitual sleep restriction (<7 h/night) is widespread in modern society and may be contributing to the development of type 2 diabetes by inducing insulin resistance in multiple tissues.

• Whole body glucose tolerance is impaired with sleep restriction but skeletal muscle (e.g., reduced peripheral insulin sensitivity) and adipose tissue (e.g., impaired insulin signaling and increased NEFA concentrations) appear to be the tissues most affected.

• Since exercise is known to improve insulin sensitivity in these tissues, it has been investigated as a therapeutic to counteract the negative ramifications of restriction sleep restriction and high intensity interval training appears to be most effective.