Excess weight is strongly linked to the development of a plethora of long-term health conditions (1 - see references at the side) and premature death (2), with significant financial implications to the National Health Service (NHS) and society (2,3). At the same time, data suggests that long-term issues with sleep are prevalent among the UK population (4) , with insufficient sleep attributed to an economic cost of 1.56% of the UK’s GDP annually (5). It is accepted that excess weight gain arises due to the influence of interacting socio-psycho-biological factors (6). The duration, quality, and timing of sleep has been found to disrupt the biological processes of circadian timing and sleep-wake homeostasis (7).
Healthy weight individuals with short sleep duration (<6 hours) have been identified as being at increased risk of a variety of metabolic indicators, including obesity (8). Furthermore, in people who habitually have short sleep duration, increasing to a healthier sleep duration of seven-to-eight hours per day has been shown to result in less weight gain and attenuated increase in fat mass compared to maintained shortsleep (9). Food intake is increased when sleep is restricted (10), potentially related to altered brain activity and levels of appetite-regulating hormones (11). Of interest, insufficient sleep has been found to have larger effect sizes on reduced health than physical inactivity (12). Nevertheless, whilst it is common practice in healthcare to address physical inactivity in an attempt to improve health and weight management, sleep issues are rarely considered or addressed.
The relationship between insufficient sleep and obesity has been considerably researched. However, the extent to which sleep duration is a limiting factor in dietary efforts to lose weight is less understood.
One prospective study followed 123 overweight and obese men and women undergoing a dietetic-led weight loss intervention for a period of 15-24 weeks (13). A greater loss of fat mass was achieved in the participants who at baseline reported longer sleep duration and better sleep quality. This observed association suggests that the outcomes from patient effort and adherence to weight loss intervention will be enhanced if sleep habits are improved. However, the study had many limitations, including the singlecohort design which opened the study to residual confounding factors that can influence sleeping behaviour and dietary intake which may have affected the identified associations. Whilst this study advanced the research focus from the well-established association between short sleep duration and obesity risk to the influence of sleeping behaviour on the success of weight loss intervention, experimental studies are required to clarify the potential mechanisms underlying the identified observation.
The impact of restricted sleep
Nedeltcheva et al (14) conducted a randomised, two-period crossover study involving ten overweight adults within a controlled clinical research environment. The study investigated the impact of recurrent restricted sleep whilst following a caloriereduced diet on body composition, energy expenditure, subjective hunger, and metabolic hormones. Participants were randomly assigned to two periods of 14 days, one of which involved spending 8.5 hours ‘time in bed’ (TIB) per night and the other spending 5.5 hours TIB. Both periods were in conjunction with a calorie restriction equivalent to 90% resting metabolic rate. Self-reported sleep at baseline was 6.5 to 8.5 hours daily. During the intervention periods, mean sleep duration was seven hours 25 minutes and five hours 14 minutes for the 8.5 hours TIB and 5.5 hours TIB elements respectively.
A key finding from this study was that, despite weight loss within the two treatment periods being similar, participants who slept <5.5 hours lost significantly less fat mass with increased fat-free mass loss compared to the 8.5 hours TIB component (14). In addition, sleep restriction was associated with higher levels of reported hunger (p = 0.043) and significantly lower resting metabolic rate (p = 0.01). Ghrelin concentration was also significantly higher (p = 0.04) after sleep restriction, identifying a potential mechanism through which the amplified changes in hunger, substrate utilisation, and metabolism arise when both sleep and calorie restriction are combined. The study’s findings outlined the importance of sufficient sleep during periods of weight loss intervention.
The use of a randomised, crossover design has both advantages and disadvantages. Whilst the potential influence of confounding variables is limited through using participants as their own control measure, the crossover design risks confounding treatment effects if ‘carry-over’ effects are not avoided by ensuring a sufficient between-treatment ‘wash-out’ period. Importantly, the study protocol implemented ensured that there was at least a three-month period between the two treatment periods (14), thus strengthening the internal validity of the study’s findings.
Nedeltcheva et al. employed a robust methodology with variables rigorously controlled within the laboratory environment (14). As a consequence, the research has a high-level of reliability and internal validity. However, the level of control implemented within the study by removing contextual ‘contaminants’ greatly affected its external validity. Ignoring the interacting, complex relationship that exists between various factors that influence health raises important questions about the ability to confidently generalise the results to larger populations in a real-world context. Whilst this study provided unique focus and insight, there was a requirement to extend the research through examining the effects of sleep restriction over a longer duration in a real-world setting.
Expanding on Nedeltcheva et al. (14) Wang et al.(15) set up an unblinded, randomised control trial conducted over an eight-week period involving 36 weight stable overweight or obese adults. Participants were randomly assigned using computer software to either the calorie restriction (control) group or calorie restriction combined with restricted sleep duration (intervention) group. A unique element of this study was the allowance for ad libitum sleep on two days of the week in order to replicate common free-living behaviour. All participants were required to follow a calorie restricted intake of 95% resting metabolic rate, with participants in the intervention group also required to restrict TIB by 90 minutes on five nights of the week. This reduction in TIB resulted in approximately 60 minutes less sleep daily on sleep restricted days over the study period.
Wang et al. corroborated Nedeltcheva et al, finding that whilst calorie restriction resulted in similar weight loss between groups, a lower percentage of fat mass and a larger percentage of fat-free mass was lost when sleep was restricted (14,15). Sleep restriction resulted in a significant decrease in leptin (p = 0.029) and increased ghrelin concentration, although for ghrelin this was not significant (p = 0.069) (15).
Wang et al. (15) aimed for dietary intake to be constant throughout the intervention so that any variation in the outcome measures between the two groups could be attributed to the sleep restriction. In order to minimise the potential for sleep restriction to impact dietary changes, the investigators provided two meals on four days of each week in addition to menu plans. It is interesting to note that, compared to baseline, a 14+/-9% and 16+/-9% calorie restriction was achieved during the intervention period for the control and intervention groups respectively. This was not as large a reduction as was originally planned. Of concern is that 30% of all participants provided <75% of their dietary intake records. As a consequence, the adherence to the planned calorie restriction and confidence in the data is questionable. The potential for variation in dietary intake between the two groups reduces the reliability of the research findings.
This randomised controlled trial (15) was at risk of Type 2 error. Although it was calculated that 36 patients were adequate to provide 80% power for the study, only 23 pre- and post-intervention serum samples (12 in control group; 11 in the intervention group) were available for assessment of ghrelin and leptin levels. The increase in ghrelin concentration found in sleep restricted participants was not found to be significant, but a larger sample size may have provided the power required to detect a significant effect, as was found by Nedeltcheva et al (14).
The impact of extending bedtimes
In an original study, Tasali et al. (16) aimed to assess whether extending the bedtimes of overweight adults who habitually have short-sleep duration impacted length of sleep and desire for food. Ten overweight adults participated in a within-participant experimental study conducted in their free-living environment over a three-week period. A two-week intervention period of extended bedtimes followed immediately after a one-week baseline period of habitual bedtimes. Education on sleep hygiene with individualised behavioural counselling was provided at the start of the intervention period. Participants were instructed to implement an extended bedtime duration of 8.5 hours. Extended bedtimes resulted in an average increase in total sleep duration of 1.5 hours. Whilst Nedeltcheva et al. found that sleep restriction was associated with higher levels of reported hunger (14), when tackling this topic from a different perspective, Tasali et al. found that extended bedtimes were associated with statistically significant reductions in both appetite (p = 0.030) and desire for unhealthy foods (p = 0.017) (16).
The findings suggest that increasing sleep duration through extended bedtimes can support dietary behaviour change. However, only subjective measures of appetite were recorded and the desirability of food was only assessed prior to breakfast. In contrast, Nedeltcheva et al. (14) measured hunger prior to every meal and at 22:30, therefore accounting for any variation in appetite throughout the day. Although statistically significant, the clinical significance of the reported 14% decrease in appetite ratings associated with increased sleep is questionable, with the study limited by the absence of objective data collection such as actual dietary intake. Whilst acknowledging that self-reported dietary consumption has a high degree of bias (17), assessing food intake would have clarified any alterations in nutrient and energy intake in relation to the reported decrease in appetite, thus identifying if this degree of appetite reduction resulted in actual dietary behaviour change and calorie reduction.
The absence of a control group and randomised design is a significant limitation of the study. Whilst striving to ensure that typical sleep-wake behaviour was not changed during the baseline period (through the blinding of participants to the planned sleep manipulation up until the start of the intervention period) aimed to use participants as their individual control measure, the nature of the pre-post study design means that factors other than the independent variable may have altered in parallel to the implementation of the intervention, thus contributing to the outcome.
The inability to exclude the influence of environmental confounding factors means that it is not possible to fully attribute detected changes in appetite and food desire to increased sleep duration. Of interest is that a relatively simple intervention such as the education and individualised behavioural counselling described by Tasali et al. supported increased sleep duration (16). Extended bedtimes were implemented and adhered to by all of the participants over the two-week intervention period. Although it is not possible to assess the sustained effects of the intervention, it was reported that participants expressed an intention to continue extended bedtimes to support longer sleep duration, suggesting perceived personal benefits in doing so. Education on sleep hygiene with individualised behavioural counselling could be incorporated as part of a holistic weight management consultation to support improved wellbeing.
Assessment and intervention in clinical practice
Assessment of sleep duration and quality via use of the Pittsburgh Sleep Quality Index (PSQI) was commonly used in the research appraised. The PSQI is a reliable and valid 19-item self-report questionnaire (18) and is appropriate for use in clinical practice (19). Certain healthcare professionals working in both primary and secondary care provide behavioural counselling on sleep hygiene within their roles. Learning from other healthcare professionals could potentially empower dietetic staff with the skills to efficiently deliver a required level of behavioural counselling. Furthermore, a number of self-help resources are available (20,21) with information provision being a low-demand intervention.
Conclusion
Insufficient sleep may compromise the benefits of dietary behaviour change. The development of metabolic and neuroendocrine changes increases the risk of reduced compliance to dietary intervention and promotion of weight regain in the future. There is a paucity of research in this area and the available literature has considerable limitations. However, the available evidence indicates that for individuals who are planning to engage in dietary behaviour change with the intention to lose weight, sleeping habits should be considered as part of a holistic, multifocused approach to weight management in order to optimise outcomes.