The case for optimising exercise timing

Optimising exercise strategies to improve metabolic disease management has long been an important area of research. Recently, investigators have sought to determine whether optimising the timing of exercise bouts is an efficacious strategy to do this^{(Reference Gabriel and Zierath2)}. While it is yet to be determined whether this will be an effective strategy on a population-wide basis, from a biological point of view there appears to be supportive evidence for this approach^{(Reference Gabriel and Zierath2)}. For example, athletes often exhibit increased strength, power, and endurance in the afternoon and evening compared to early morning^{(Reference Chtourou and Souissi12)}. More world records are broken by athletes competing in the early evening, even when controlling for environmental conditions and scheduling biases^{(Reference Atkinson and Reilly13)}. Disruptions in circadian rhythms can also harm athletic performance. Eastward trans-meridian travel negatively affects intermittent sprint performance and psychological indicators of fatigue more than westward travel^{(Reference Fowler, Knez and Crowcroft14)}. Further, the chronotype (whether they tend towards ‘morning-ness’ or ‘evening-ness’) of an athlete can influence exercise performance at various times of the day^{(Reference Facer-Childs and Brandstaetter15)}. When looking at cellular physiology, one of the main mechanisms by which exercise has beneficial health outcomes is by remodelling skeletal muscle to improve mitochondrial oxidation, for example by increasing total activity of enzymes such as citrate synthase^{(Reference Leek, Mudaliar, Henry, Mathieu-Costello and Richardson16)} and therefore improving lipid and glucose oxidation^{(Reference Gabriel, Al-Tarrah and Alhindi17,Reference Alhindi, Vaanholt and Al-Tarrah18)} . However, many of the improved substrate-handling effects of exercise appear to be acute, with studies suggesting that some beneficial effects of acute exercise disappear within 48 hours^{(Reference Gabriel, Pugh, Pruneta-Deloche, Moulin, Ratkevicius and Gray19)}. Given this relatively short timeline of exercise efficacy, it appears to be worth considering how to optimise exercise bouts around other events such as nutritional and medicine intake to maximise beneficial outcomes. Another aspect that makes exercise timing worthy of study is the circadian rhythm of the molecular systems that exercise beneficially remodels. For example, mitochondrial function has a diurnal rhythm in healthy skeletal muscle, with lowest adenosine diphosphate (ADP)-stimulated mitochondrial respiration at 1 PM and highest at 11 PM^{(Reference van Moorsel, Hansen and Havekes20)}. However, mitochondrial rhythm is disrupted in skeletal muscle myocytes from people with type 2 diabetes^{(Reference Gabriel, Altintaş and Smith21)}. Additionally, mitochondrial genes and function appear to be partially under the control of the molecular clock, with evidence for bi-directional signalling^{(Reference Gabriel, Altintaş and Smith21,Reference Lassiter, Sjögren, Gabriel, Krook and Zierath22)} . Given the clear circadian rhythmicity of mitochondrial function in skeletal muscle, one may ask if it is possible to ‘re-set’ the circadian rhythm of mitochondria in the skeletal muscle of people with type 2 diabetes by using precision timing of exercise.

Several studies now demonstrate divergent time-of-day exercise effects on glycaemia in people living with either type 2 diabetes or obesity^{(Reference Mancilla, Brouwers, Schrauwen-Hinderling, Hesselink, Hoeks and Schrauwen23–Reference van der Velde, Boone and Winters-van Eekelen26)}. However, there are far fewer studies testing time-of-day exercise outcomes on glycaemia and other metabolic outcomes in metabolically healthy individuals. Thus, there is not yet evidence to support a strong divergence between morning and afternoon/evening exercise on glycaemia in metabolically healthy individuals^{(Reference Tanaka, Ogata and Park27)}. However, one area of circadian biology that appears to have divergent outcomes in time-of-day exercise studies in metabolically healthy individuals is sleep^{(Reference Driver and Taylor28)}. Sleep is crucial to metabolic health and disruptions in sleep have consistently been linked to poorer metabolic health outcomes^{(Reference Vetter, Dashti and Lane29–Reference Giannos, Prokopidis and Candow31)}. These positive effects of exercise on sleep could be useful for addressing disrupted sleep patterns caused by medical conditions, seasonal changes in daylight, or shift work. However, exercise at inappropriate times might have adverse effects. For instance, the benefits of intense exercise might be reduced if done at a time the individual is not accustomed to. Additionally, chronotype may play a role in sensitivity to exercise at different times of day for sleep outcomes. For example, adolescent athletes with an evening chronotype who performed exercise in the evening saw little sleep disruption when assessed with polysomnography^{(Reference Saidi, Peyrel and del Sordo32)}. Whereas those with a morning chronotype saw a significant disruption to their sleep architecture at the same time of evening exercise, compared to no sleep disruption with midday exercise^{(Reference Saidi, Peyrel and del Sordo32)}. This study^{(Reference Saidi, Peyrel and del Sordo32)} included a mixed-intensity exercise bout; however, studies assessing diurnal exercise outcomes have used a variety of exercise intensities and modalities^{(Reference Egan and Zierath33)}. Different exercise intensities lead to varying metabolic and signalling outcomes, including the recruitment of specific types of skeletal muscle fibres. High-intensity or resistance exercises primarily recruit type II fibres, which are fast-twitch, fatigable, and more glycolytic^{(Reference Egan and Zierath33)}. In contrast, low-intensity endurance exercises predominantly engage type I fibres, which are highly oxidative^{(Reference Egan and Zierath33)}. The canonical circadian biology in each muscle fibre-type shows a similar circadian rhythmicity. However, each fibre-type exhibits unique expression of non-canonical, but diurnally cycling biological processes^{(Reference Dyar, Ciciliot and Tagliazucchi34)}. The differential recruitment of fibre types during exercise may thus influence circadian gene expression in an exercise-specific manner. One example of this phenomenon is that varying exercise intensities can have differential effects on postprandial triglyceride metabolism throughout the following day^{(Reference Gabriel, Pugh, Pruneta-Deloche, Moulin, Ratkevicius and Gray19)}. Therefore, adjusting the timing of exercise interventions to mitigate spikes in postprandial metabolites could be a preventive strategy against metabolic diseases. Additionally, a burgeoning area of research focuses on the interplay between periodised nutrition and exercise responses^{(Reference Hawley, Lundby, Cotter and Burke35)}. Many periodised nutrition protocols involve altering carbohydrate availability before, during, or after exercise sessions^{(Reference Hawley, Lundby, Cotter and Burke35,Reference Marquet, Brisswalter and Louis36)} . For instance, performing an intense workout in the evening followed by a low carbohydrate intake, which lowers muscle and liver glycogen levels, and then sleeping, has shown benefits in exercise performance and the skeletal muscle signalling response of the lipid oxidation pathway^{(Reference Hawley, Lundby, Cotter and Burke35,Reference Marquet, Brisswalter and Louis36)} . Given the key role that nutritionally derived substrates play in exercise capacity and the health benefits derived from exercise training, it is important to consider nutritional timing in relation to the circadian biology of exercise and its role in managing type 2 diabetes.

The importance of nutritional timing

Possibly the most important aspect of disease management in type 2 diabetes is the regulation of glycaemia within optimal ranges. Clearly, nutritional intake is highly important in this. Recently, studies have shown possible benefits of considering timing of nutritional intake in managing glycaemic regulation. Diurnal changes in glucose tolerance have been recognised in human subjects for many years^{(Reference Van Cauter, Polonsky and Scheen37)} where sensitivity to elevated glucose concentration is greatest in the early morning and then declines over the course of the day. This daily change is not dependent upon changes in gastrointestinal function, but instead appears to be the result of altered glucose utilisation and insulin sensitivity, with maximal insulin sensitivity occurring in the early morning and decreasing throughout the day^{(Reference Johnston38)}.

Recent progress in the field of circadian biology suggests that not only what we eat but when we eat influences our physiology and metabolism and thereby affects our health^{(Reference Panda39)}. The time of day when food is ingested affects body weight, body composition, glucose regulation, lipid homeostasis, the gut microbiome, cardiac function and inflammation^{(Reference Chaix, Lin, Le, Chang and Panda40)}. Weight management is an integral part of reducing cardiometabolic risk; and it is known that modest amounts of weight loss can improve glycaemic control. The question remains, whether time of feeding can influence metabolism in the absence of overall energy restriction since understanding the mechanisms of dietary influence on circadian physiology could open new approaches for health management. For example, if time of eating can improve metabolic control in the absence of overall energy restriction, this would provide an additional dietary strategy that could be exploited against the rising prevalence in type 2 diabetes. Data from studies in mice show that mice fed a high-fat diet ad libitum, albeit within a time-limited feeding period (8-12 hours in 24h), have reduced adiposity and liver steatosis, as well as improved glucose tolerance and reduced cholesterol levels, compared with mice fed ad libitum over 24hrs; importantly, these improvements occur in the absence of changes in energy intake^{(Reference Chaix, Lin, Le, Chang and Panda40)}. This approach to eating is called, ‘time restricted feeding’ (TRF). In TRF, periods of normal energy intake are punctuated by periods of energy restriction or fasting with typically around 8:16 hours for feeding and fasting, respectively^{(Reference Varady, Bhutani, Church and Klempel41)}. This contrasts with the usual daily dietary pattern for humans^{(Reference Gill and Panda42)}, where the eating period is often up to 14hr with 10hr (or less) of fasting during sleep. This research suggests that timing of eating has an important influence on indices of metabolic health, independent of weight loss. Since much of the empirical work is from mouse models^{(Reference Hatori, Vollmers and Zarrinpar43)}, it may not translate directly to humans. While several uncontrolled studies in humans have shown that TRF reduces body weight by a modest 1–3% over a period of 2 to 16 weeks^{(Reference Gill and Panda42,Reference Gabel, Hoddy and Haggerty44,Reference Tinsley, Moore and Graybeal45)} on the basis of the study designs it is not possible to conclude that TRF per se underpinned the weight loss reported due to the lack of a control group. Interestingly, when energy restriction is absent, or minimal, current evidence suggests that time of eating can influence glucose metabolism necessitating further study of the optimal feed-fasting window in overweight subjects. There has been little human work published in this area. Two studies have investigated TRF in males only with prediabetes (metabolic syndrome). One study involved 15 participants conducted as a cross-over trial, with 7 days per treatment arm^{(Reference Hutchison, Regmi and Manoogian46)}, applying a 9hr eating window. Advice was given to eat normal diet either as eTRF (morning eating, 8am–5pm) or dTRF (delayed eating, 12noon–9pm). This study was conducted in a free-living population, and there was no attempt to standardise food intake, with no food intake data reported. Subjects lost a small amount of weight on both arms (−0.8kg on eTRF and −1.3kg on dTRF), with time of eating compliance monitored using continuous glucose monitoring. There was a decreased incremental area under the curve (iAUC) in glucose in both TRF treatments, with a greater improvement with the morning feeding regime (−36% eTRF & −21% dTRF). The improvement in glycaemic control was not explained by changes in gastric emptying or gastrointestinal hormone release, although there was a tendency to reduce postprandial insulin in both TRF treatments. Mean fasting glucose by continuous glucose monitoring was lower only during the eTRF treatment versus baseline, with no difference between postprandial fed glucose concentration. In a second study by Sutton et al (2018), controlled diets were applied over 5 weeks in males only and improvements in glucose/insulin metabolism were reported in the absence of weight loss, with subjects keeping weight stable. These improvements were linked to early feeding as eTRF (early eating, 18hr fast) and dTRF (12 hr eating & 12 hr fast). In response to a 3hr oral glucose tolerance test (OGTT), TRF improved postprandial insulin, insulin sensitivity and β-cell function after 5 weeks, while the blood glucose response remained unchanged. This study only applied OGTT to capture the morning postprandial response. The traditional eating pattern of many Western societies of having the largest energy load in the evening (and either no, or only a small, breakfast) may contribute to the metabolic syndrome through deterioration in postprandial glucose mechanisms and insulin sensitivity. There is a known circadian impact of meal timing, with poorer glucose tolerance at night despite identical meals and equidistant fasting lengths^{(Reference Morgan, Aspostolakou, Wright and Gama47,Reference Grant, Ftouni and Nijagal48)} . Unlike animal models, conducting circadian (or diurnal) studies in humans is challenging practically and teasing out the circadian and diet influences on peripheral circadian clocks are difficult without invasive testing.

Part of management of type 2 diabetes, is achieving optimal weight, for height, particularly for those with abdominal obesity. Although dietary management is considered central to combatting obesity and metabolic diseases, the emphasis has so far been on caloric reduction and exercise⁽⁴⁹⁾ to reduce co-morbidity impact. We suggest that an adjunct to this approach is optimal timing of meals. Diurnal changes in glucose tolerance have been recognised in human participants for many years^{(Reference Van Cauter, Polonsky and Scheen50)} where sensitivity to elevated glucose concentration is greatest in the early morning and then declines over the course of the day. This daily change is not dependent upon changes in gastrointestinal function, but instead appears to be the result of altered glucose utilisation and insulin sensitivity, with maximal insulin sensitivity occurring in the early morning and decreasing throughout the day^{(Reference Johnston38)}. Although the mechanism for the diurnal variation in insulin sensitivity is undefined, some studies have explored this phenomenon^{(Reference Morgan, Aspostolakou, Wright and Gama47,Reference Wehrens, Christou and Isherwood51)} , and one notable acute (within day) feeding trial supports an approach of optimising meal timing^{(Reference Morgan, Shi, Hampton and Frost52)}. This study assessed glucose response to early and late feeding in healthy subjects. The data indicate that early morning feeding in combination with low-glycaemic index (GI) foods can improve glucose tolerance and insulin secretion, when compared to evening eating and high-GI foods. Our own recent work has applied controlled experimental design to tease out the effect of time of day and eating patterns on energy balance. We explored the timing of total caloric intake interaction with incretin hormones to alter satiety^{(Reference Ruddick-Collins, Johnston, Morgan and Johnstone53)}. Our research demonstrates that in a within-subject cross-over trial, people living with overweight who ate more of their calories in the morning had better satiety on a low-calorie diet compared to when they ate the majority of their calories in the evening (isocaloric over 24 hours)^{(Reference Ruddick-Collins, Morgan and Fyfe54)}. Both diets produced similar weight loss, with no differences in energy metabolism. The morning-loaded energy intake resulted in lower daily subjective appetite and hunger. Further, we suggest that in a free-living setting, that this would support behavioural changes to achieve weight loss.

Diurnal timings of exercise, nutrition and interaction with medicine

When considering the precision timing of exercise and nutrition and their application to disease management, it is imperative to consider the holistic diurnal environment. Many people with type 2 diabetes take a daily or twice daily dose of metformin as part of their disease management strategy. Further, many people with gestational diabetes (GDM) (affects 1 in 20 pregnancies in the UK) and polycystic ovary syndrome (1 in 10 women in the UK) are also prescribed metformin (>85% of women with GDM are prescribed metformin) in addition to concomitant clinical lifestyle recommendations. As stated above, a central mechanism by which exercise improves whole-body glycaemic regulation is through improving glucose uptake, glucose metabolism and mitochondrial function in skeletal muscle. Interestingly, metformin appears to modulate glycaemia independent of skeletal muscle, instead acting on gut, liver and kidney tissues to improve glycaemia^{(Reference Tobar, Rocha and Santos55,Reference Foretz, Guigas and Viollet56)} . Evidence also shows that metformin has a strong circadian profile in plasma^{(Reference Türk, Scherer and Selzer57)}. Given the tissue-specific nature of metformin action, seeking a chrono-medicine approach to augment physiological delivery of metformin to organs by which metformin can beneficially regulate glycaemia, and reducing accumulation of metformin in skeletal muscle near the time of exercise may be beneficial. Indeed, our recent study^{(Reference Carrillo, Cope and Gurel58)} shows that morning exercise significantly reduces glycaemia in individuals with T2D who are also taking metformin, whereas evening exercise had no noticeable impact on glycaemic levels. The reduction of glycaemia in the morning exercise trial was driven by people who consumed metformin before breakfast, rather than after breakfast. Morning exercise combined with pre-breakfast metformin both acutely and persistently reduced area under the curve (AUC) glucose compared to morning exercise combined with post-breakfast metformin until the final week (week 6) of the intervention^{(Reference Carrillo, Cope and Gurel58)}. These findings have parallels to previous data that suggest consuming metformin before a meal improved efficacy of glycaemic regulation^{(Reference Hashimoto, Tanaka and Okada59)}. This effect on glycaemia may be partly due to the pharmacokinetic interaction of metformin with meal intake and intrinsic circadian rhythms. For instance, when an 850 mg tablet of metformin is taken with food, its bioavailability is 24% lower and its peak concentration is delayed by about 37 minutes compared to the fasting state^{(Reference Sambol, Brookes and Chiang60)}. Additionally, metformin’s pharmacology significantly depends on the time of day in humans, which may be linked to factors like glomerular filtration rate, renal plasma flow and renal organic cation transporter 2 activity^{(Reference Türk, Scherer and Selzer57)}. Although metformin has a relatively long half-life in the blood (approximately 18 hours depending on dose/method^{(Reference Larsen, Rabol, Hansen, Madsbad, Helge and Dela61)}, less is known about its accumulation in skeletal muscle, a key organ in the response to exercise. It is known that concurrent exercise alters the pharmacokinetics of acute metformin administration^{(Reference Kristensen, Lillelund and Kjobsted9,Reference Larsen, Rabol, Hansen, Madsbad, Helge and Dela61)} , with apparently varying concentrations of metformin in skeletal muscle before, during and after exercise^{(Reference Kristensen, Lillelund and Kjøbsted62)}. Therefore, it is plausible that the timing of exercise and metformin intake may interact to influence the pharmacokinetics and glycaemic-modulating effects of metformin (Figure 1).

Figure 1. Schematic of chrono-medicine considerations in management of metabolic disease. Morning and evening exercise may have a substantially different effect on 24-hour glycaemia^{(Reference Mancilla, Brouwers, Schrauwen-Hinderling, Hesselink, Hoeks and Schrauwen23–Reference van der Velde, Boone and Winters-van Eekelen26)}. Higher-intensity exercise may exacerbate diurnal exercise outcomes^{(Reference Gabriel and Zierath2)}. A diet with morning-loaded calories may increase satiety relative to an isocaloric diet with calories-loaded in the evening^{(Reference Ruddick-Collins, Morgan and Fyfe54)}. Restricting food intake time during the day may improve metabolic health and reduce overall calorie consumption^{(Reference Varady, Bhutani, Church and Klempel41)}. Recent research suggests that integrating precision-timed medicine, exercise and nutrition may be an important strategy for metabolic disease management^{(Reference Carrillo, Cope and Gurel58)}.

Summary

In summary, chrono-medicine is an important emerging research area that has a great deal of promise as a cost-effective strategy for metabolic disease management. Our research indicates that it may be possible to optimise the timing recommendations for concomitant exercise and metformin treatment. Specifically, our findings suggest that morning exercise combined with pre-breakfast metformin intake may benefit the management of glycaemia in people with T2D. Additionally, nutritional studies have demonstrated the potential of using precision timing in nutritional intervention studies. To translate these approaches into well-evidenced healthcare strategies, researchers must consider the holistic diurnal environment including nutritional intake, physical activity, medicine intake, sleep and occupation. Only when using this holistic approach will we be able to take steps towards a more established translational pathway for chrono-medicine approaches in metabolic disease management.