Examining the Role of Exercise Timing in Weight Management: A Review

Abstract

Many adults cite exercise as a primary strategy for losing weight yet exercise alone is modestly effective for weight loss and results in variable weight loss responses. It is possible that some of the variability in weight loss may be explained by the time of day that exercise is performed. Few studies have directly compared the effects of exercise performed at different times of the day (i.e., morning versus evening exercise). Results from these existing studies are mixed with some studies demonstrating superior weight and fat mass loss from morning exercise, while other studies have found that evening exercise may be better for weight management. Exercise timing may alter modifiable lifestyle behaviors involved in weight management, such as non-exercise physical activity, energy intake, and sleep. The purpose of this review is to summarize evidence for and against time-of-day dependent effects of exercise on weight management. Although limited, we also review studies that have examined the effect of exercise timing on other lifestyle behaviors linked to body weight regulation. While exercise at any time of day is beneficial for health, understanding whether there is an optimal time of day to exercise may advance personalized treatment paradigms for weight management.

INTRODUCTION

Despite significant efforts to thwart the increasing prevalence of overweight and obesity (BMI ≥25.0 kg/m²), rates have continued to increase amongst U.S. adults [1–3]. It has been projected that by 2030 nearly 50% of US adults will have obesity (BMI ≥30 kg/m²) and nearly 25% will have severe obesity (BMI ≥35 kg/m²) [4]. Obesity is associated with increased risk of cardiovascular disease [5, 6], type 2 diabetes [7, 8], cancer [9–11], all-cause mortality [12], and lower quality of life [13]. Exercise and physical activity (PA) are critically important to prevent and treat overweight and obesity. Current guidelines recommend high levels of PA for weight management [14, 15]; however, there is considerable unexplained individual variability in the weight loss response to exercise [16–19].

Recent evidence suggests that circadian rhythms (i.e., endogenous rhythms that occur with near 24h rhythmicity without external stimuli) and timing of lifestyle behaviors influence body weight regulation [20–23]. For example, evidence from observational and experimental studies have suggested that the timing of food intake may influence bodyweight [24–33]. Most evidence in humans suggests that a greater proportion of food intake consumed earlier in the day helps to promote a healthier, lower bodyweight [25, 26, 31–33]. There have been far fewer studies examining the effect of exercise timing (i.e., the time of day of exercise) on body weight regulation; however, this is an emerging area of research. The purpose of this review is to provide an overview of the current evidence on the effect of exercise timing on body weight. In addition, we examine three potential areas by which exercise timing may affect body weight: 1) energy intake (EI), 2) energy expenditure (EE) and substrate metabolism, 3) and sleep. The conceptual framework for how exercise timing may affect these factors and ultimately weight management is illustrated in Figure 1. Finally, we address the strengths and limitations of current studies, recognize research gaps, and identify future areas of research that are needed.

Figure 1: Conceptual framework for the differential impact of exercise timing on weight loss and weight management.

Exercise is a key component of lifestyle changes to promote weight loss and manage healthy body weights. Exercise has systemic impacts on factors related to weight management including energy balance, energy expenditure/substrate oxidation, and sleep. The effect of morning (yellow) vs. evening (blue) exercise on these components is likely different and illustrated in the conceptual framework. Solid lines represent well supported relationships whereas dotted lines represent proposed relationships.

EFFECT OF EXERCISE TYPE ON BODY WEIGHT AND BODY COMPOSITION

Exercise and PA are important behaviors for achieving and maintaining a healthy body weight [14]. Prior studies have demonstrated that weight loss from exercise without dietary changes results in modest weight loss [18, 34–36]. However, with high doses of exercise (e.g., >2000 kcal/week), clinically meaningful weight loss is attainable [37–39]. Aerobic exercise training is considered the most effective form of exercise to reduce weight and fat mass [14]. In contrast, resistance exercise training may be beneficial for retaining lean mass and may elicit modest reductions in fat mass when performed in combination with aerobic exercise or diet restriction [14]. Current guidelines and recommendations related to weight management primarily focus on aerobic exercise [14, 15, 40, 41]. Therefore, for the purpose of this review, we focus on the impact of aerobic exercise performed at different times of day on body weight and body composition.

THE VARIOUS DEFINITIONS OF EXERCISE TIMING

To date, there are no standard definitions of exercise timing. This is largely due to the fact that timing of exercise can be anchored to clock time, markers of circadian phase (e.g., dim light melatonin onset, core body temperature), habitual sleep and wake times, and/or eating events (e.g., pre vs. post meal; fasted vs. non-fasted). Mechanistic studies focusing on physiological pathways have traditionally defined exercise timing relative to circadian time and eating events, while secondary analyses and pragmatic clinical trials have defined exercise timing relative to clock times. All of these definitions are relevant to bodyweight regulation and the current review; thus, we will review data from several studies with different definitions of exercise timing. However, the focus of this review will be on morning vs. evening exercise because these are the most common exercise times and are generalizable to most adults. The definitions of morning and evening exercise vary by study and are reported as such below.

EFFECT OF EXERCISE TIMING ON BODY WEIGHT AND BODY COMPOSITION

There is accumulating evidence from observational and clinical studies that the timing of exercise and PA affects body weight regulation [42–49]. Some evidence suggests that morning exercise and PA is associated with more favorable changes in weight and body composition [42, 45] and increased ability to prevent weight regain [46, 47, 49]. However, some data suggest that evening exercise may be more beneficial for reducing fat [44, 48] (Table 1). Due to a paucity of direct head-to-head comparisons of different exercise times of day, it remains difficult to discern whether there is an optimal time of day for exercise to maximize weight loss benefits. The evidence to date, summarized below, is from observational studies, small randomized trials, and retrospective analyses [42–50].

Table 1.

Summary of Studies on the Effect of Exercise Timing on Body Weight

Reference	Sample Size	Participants	Study Design	Intervention	Study Duration	Weight Outcomes	Related Outcomes
Alizadeh et al. [42]	48	Inactive females aged 20–45 y with a BMI ≥25.0–29.9 kg/m2	RT	Frequency: 3 d/wk Intensity: at ventilatory threshold Type: Supervised treadmill running Time: 30 min/d AM-Ex: 08:00 – 10:00 PM-Ex: 14:00 – 16:00	6 weeks	*Δ in Weight: AM-Ex = −1.6 kg PM-Ex = −0.3 kg No differences in change in body composition	Δ in EI: AM-Ex = −362 kcal/d PM-Ex = −28 kcal/d 77% retention; adherence to exercise not reported
Willis et al. [45]	88	Inactive males and females aged 20–45 y with a BMI ≥25.0 – 40.0 kg/m2	NR	Frequency: 5 d/wk Intensity: 70–80% MHR Type: Primarily supervised treadmill walking/running Time: varied by participant as prescriptions were based on EE; 400 and 500 kcals per session AM-Ex: 07:00 – 11:59 PM-Ex: 15:00 – 19:00 Control: No-exercise	10 months	*Δ in Weight: AM-Ex = −7.2 ± 1.2% PM-Ex = −2.1 ± 1.0% Control = 0.5 ± 1.0% Changes in fat mass followed the same trends at weight loss	Δ in EI: AM-Ex = −63 ± 444 kcal/d PM-Ex = 121 ± 484 kcal/d Δ in Non-Exercise EE: AM-Ex = 28 ± 446 kcal/d PM-Ex = −105 ± 510 kcal/d
Creasy et al. [49]	90	Cohort of weight loss maintainers compared to control with and without obesity	OBS	-	-	WLM = BMI of 23.7 ± 2.4 kg/m2; maintaining weight loss for 9.5 ± 10.2 years NC = BMI of 22.7 ± 2.0 kg/m2 OC = BMI of 33.4 ± 5.1 kg/m2	*MVPA 0–3 hours after waking up: WLM: 25.1 ± 23.1 min NC: 14.3 ± 12.3 min OC:12.3 ± 14.3 min
Schumacher et al. [46]	375	Cohort of weight loss maintainers; mostly white females with a mean age of 53.4 y	OBS	-	-	WLM = Current BMI of 26.5 ± 4.8 kg/m2; maintaining a significant weight loss	WLM with temporally consistent physical activity patterns engaged in the higher amounts of MVPA compared to WLM with temporally inconsistent patterns (350 min/wk vs. 285 min/wk; p=0.03) WLM with temporally consistent patterns predominantly were active in the early mornings (4:00–8:59).
Chomistek et al. [47]	7,157	Cohort of women in the Women’s Health Study; mean BMI = 26.1 ± 4.9 kg/m2	OBS	-	-	Women in the lowest quartile of accelerometer counts prior to 12:00 (<39% of total counts) have 26% higher odds of having obesity compared to those in the highest quartile (≥54% of total counts).	-
Di Blasio et al. [44]	42	Inactive post-menopausal women aged 48–58 y with a BMI ≥25.0 kg/m2	NR	Frequency: 4 d/wk Intensity: 55% HRR Type: Walking Time: 50 min/d Morning Exercise: 07:00 – 09:00 Evening Exercise: 18:00 – 20:00	3 months	Δ in Weight: AM-Ex: 0.2 ± 2.0 kg PM-Ex: −0.02 ± 1.7 kg *Δ in FM: AM-Ex:−0.2 ± 1.5 kg PM-Ex: −1.7 ± 2.4 kg	Adherence to Exercise: AM-Ex: 83.2% PM-Ex: 87% Δ in EI: AM-Ex: −40 ± 209 kcal/d PM-Ex: 101 ± 285 kcal/d *Δ in Meal-timing: AM-Ex: 0.4 ± 8.1% increase in AM intake PM-Ex: 6.1 ± 5.7% increase in AM intake
Marinac et al. [50]	125	US Adults with a mean age of 35.3±11 y and a mean BMI of 27.6±6.5 kg/m2	OBS	-	-	Timing of physical activity (morning=prior to 12:00 vs. evening = after 19:00) was not associated with BMI.	-
Mancilla et al. [48]	32	Inactive males age 58±7 y at risk for or diagnosed with type 2 diabetes and a BMI ≥26.0 kg/m2	NR	Frequency: 3 d/wk (aerobic=2 days; resistance=1 day) Intensity: aerobic=70% Wmax Resistance=60% of MVC Type: aerobic=cycling resistance=3×10 reps of leg extension, leg press, chest press, lat pull down, tricep and bicep curls, abdominal crunches, horizontal row Time: aerobic=30 min/d; resistance=until completion Morning Exercise: 08:00 – 10:00 Evening Exercise: 15:00 – 18:00	12 weeks	Δ in Weight: AM-Ex: 0.7 ± 1.6 kg PM-Ex: −0.6 ± 2.5 kg *Δ in FM: AM-Ex: −0.2 ± 1.0 kg PM-Ex: −1.2 ± 1.3 kg	Adherence to Exercise: Overall= 98% compliance Δ in VO2max: AM-Ex: 1.8 ± 2.2 ml/kg/min PM-Ex: 3.0 ± 2.1 ml/kg/min

Abbreviations and symbols: BMI: body mass index, y: years, d/wk: days per week, min/d: minutes per day, RT: randomized trial, OBS: observational, NR: not-randomized, AM-EX: morning exercise, PM-Ex: evening exercise, MHR: maximal heart rate, HRR: heart rate reserve, Δ: change, BF: body fat, FM: fat mass, EI: energy intake, EE: energy expenditure, MVPA: moderate to vigorous physical activity, WLM: weight loss maintainers, NC: controls without obesity, OC: controls with obesity, US: United States, Wmax: maximum workload, MVC: maximal voluntary contraction,

^*signifies a significant difference (p<0.05)

Results from observational studies suggest that there may be an effect of time of day of PA on body weight and BMI. Chomistek et al. examined the PA patterns of 7157 women enrolled in the Women’s Health Study. Women who engaged in the highest amount of PA prior to 12:00 had 26% lower odds of having obesity compared to those who engaged in the lowest amount of PA prior to 12:00 [47]. In contrast, Marinac et al. found that neither engaging in morning PA (prior to 12:00) nor evening PA (after 19:00) was associated with BMI in a convenience sample of 125 US adults [50]. In addition to timing of PA, the authors also examined the association between two phenotypes (“early bird” vs. “night owl”) and BMI. “Early birds” and “night owls” were defined by a combination of factors including the timing of last outdoor light exposure, last indoor light exposure, first indoor light exposure, bedtime, waketime, and timing of last meal. The “night owl” phenotype was associated with a lower BMI; however, this was not significant after adjustment for covariates. While these studies do not provide congruent results, they highlight the need for further investigation in this area.

There have been very few randomized control trials comparing different exercise times of day on changes in body weight and body composition. Alizadeh et al. [42] randomized women (age 20–45 years, BMI of 25–30 kg/m²) to six weeks of supervised morning (08:00–10:00) or afternoon (14:00–16:00) exercise of equivalent duration (90 min/wk). While this study was not powered to detect differences in weight change, in a completers analysis, body weight decreased more in morning exercisers compared to afternoon exercisers (−1.6 vs. −0.3 kg, SD not reported, p=0.04 for group*time interaction). This study had several strengths including the randomized design, provision of supervised exercise, and measurement of body composition and EI. However, the sample size was relatively small (n=48), and the population studied was exclusively women with overweight between the ages of 20–45 years. Thus, the results of this study may not be generalizable to other populations such as men, postmenopausal women, and individuals with obesity (BMI ≥30.0 kg/m²). Moreover, the exercise intervention was modest (3 days per week, 30 minutes per session of treadmill running at the individualized ventilatory threshold), exercise adherence was not reported, and the intervention duration was very short (6 weeks). This level of exercise is not expected to result in clinically significant weight loss [14]; thus, it is not surprising that average weight loss was minimal (<2 kg). Further, it is unlikely that the differences in weight loss observed between morning exercise and evening exercise would result in any clinically meaningful differences in other health-related outcomes.

Willis et al. retrospectively examined the effect of morning exercise vs. evening exercise on weight loss and energy balance in the Midwest Exercise Trial 2 (MET-2) [45]. MET-2 was a highly controlled 10-month exercise-only intervention which examined the effects of 2000 and 3000 kcal/wk of supervised exercise (5 days per week) in men and women (18–30 years, BMI 25–40 kg/m²). Importantly, only individuals who were >90% adherent to the exercise intervention were retained in MET-2. In this post-hoc analysis, participants were categorized based on the time of day in which they completed most of their exercise sessions: morning exercisers: >50% of sessions completed between 7:00–11:59; (n = 21, mean ± SD; 70 ± 14% of exercise sessions completed in the morning), evening exercisers: >50% of sessions completed between 15:00–19:00; (n = 25, 66 ± 12% of exercise sessions completed in the evening), and sporadic exercisers: <50% of sessions completed in any time category; (n = 24). Morning exercisers lost significantly more weight (−6.5 ± 5.3 kg, −7.2%) compared to evening exercisers (−2.2 ± 4.5 kg, −2.2%) at 10 months despite similar levels of exercise EE and similar baseline characteristics. Sporadic exercisers exhibited weight loss of −5.0 ± 5.3 kg (−5.5%) of bodyweight which was intermediate between morning and evening exercisers. Sensitivity analyses confirmed that morning exercisers achieved superior weight loss compared to evening exercisers regardless of randomization group (2000 kcal/wk group vs. 3000 kcal/wk group) or sex. In addition, morning exercisers lost significantly more fat mass (−6.2 ± 1.1 kg) compared to evening exercisers (−1.6 ± 0.9 kg) as measured using dual-energy x-ray absorptiometry (p=0.008). Strengths of this study include fully supervised exercise with a prescription based on exercise EE measured and adjusted throughout the study, clinically relevant study duration, and inclusion of men and women. However, it is important to note that participants were not randomized to morning or evening exercise in this secondary analysis and as a result, unobserved confounders may have impacted these findings. Additionally, this study utilized arbitrary ranges for morning (07:00–11:59) and evening exercise (15:00–19:00); these defined clock times may not appropriately reflect morning and evening for all people.

Not all studies show a benefit of morning exercise for weight loss. Di Blasio et al. performed an interventional study to examine the effect of morning exercise vs. evening exercise on changes in weight, body composition, and eating behavior [44]. Forty-two postmenopausal women self-selected either morning (7:00–9:00) or evening exercise (18:00–20:00) for 3 months. The exercise intervention consisted of 4 sessions per week of walking for 50 minutes per session at an intensity equivalent to 55% of heart rate reserve (200 minutes per week). Two sessions per week were supervised by a trainer and two sessions per week were unsupervised. Overall exercise adherence was similar between groups (83.2% for morning and 87.0% for evening exercise). In contrast to the previously described studies, weight change was similar between morning exercise (0.2 ± 2.0 kg) and evening exercise (−0.02 ± 1.7 kg); however, evening exercise exhibited a greater decrease in fat mass (−1.7 ± 2.4 kg) compared to morning exercise (−0.2 ± 1.5 kg), as measured using bioelectrical impedance (p=0.037). There were several strengths of this study including high levels of adherence to the exercise intervention, partial exercise supervision, and well controlled exercise intensity. However, because participants were allowed to self-select the timing of exercise, it is difficult to determine if the differential weight loss was due to time of day or other confounders.

Another recent secondary analysis also found that evening exercise may lead to more favorable effects on body weight, body composition, and insulin sensitivity in men who were at risk for or diagnosed with type 2 diabetes [48]. Thirty-two men self-selected to exercise in the morning (08:00–10:00) or the early evening (15:00–18:00) for 12 weeks. All exercise was supervised and included two days per week of cycling at 70% of maximum workload for 30 minutes and 1 day per week of resistance training exercises (3 sets of 10 repetitions) at 60% of maximum voluntary contraction. Overall compliance to exercise training was 98%. Changes in body weight were not statistically different between morning exercise (0.7 ± 1.6 kg) and early evening exercise (−0.6 ± 2.5); however, early evening exercise had greater reductions in body fat (−1.2 ± 1.3 kg) compared to morning exercise (−0.2 ± 1.0 kg, p=0.03). In addition, early evening exercise resulted in superior improvements in peripheral insulin sensitivity, insulin-mediated suppression of adipose tissue lipolysis, and morning fasting glucose levels. There were several strengths of this study including high levels of adherence to the exercise intervention, inclusion of aerobic and resistance exercise, exercise supervision, and well controlled exercise intensity. However, the non-randomized design makes it difficult to discern whether the findings were due to time of day of exercise or other potential confounders.

Prescribing exercise based on time of day is a promising future direction of overweight and obesity treatments. A recent rigorously designed randomized control trial by Brooker et al. [43] assessed the feasibility, safety, and acceptability of prescribing morning versus evening exercise in 20 adults with overweight and obesity (BMI: 25–40 kg/m²). Participants (age 18–60 years) were randomized to 12 weeks of morning exercise (06:00–9:00, n=9), evening exercise (16:00–19:00, n=7), or a no intervention control (n=4). Exercise volume was 250 minutes per week at moderate intensity. The intervention started as fully supervised (5 days per week, 50 minutes per session) and was tapered to two sessions per week of supervision over the intervention. Both morning exercise and evening exercise resulted in favorable changes in BMI and percent body fat compared to control. Because of the limited sample size, direct comparisons of weight outcomes between morning and evening exercise were not performed. The study did report high adherence to supervised exercise sessions at both times of day (94% in morning exercise, 87% in evening exercise). These data indicate that prescribing exercise based on time of day is feasible and acceptable in the setting of overweight and obesity treatments. A future randomized control trial that is powered to detect differences in body composition between morning and evening exercise is warranted.

Timing of exercise and PA may also play a role in the maintenance of weight loss. In the National Weight Control Registry Ostendorf et al. found that weight loss maintainers (individuals who maintained a weight loss of >13.6 kg for >1 year) engaged in higher volumes of moderate to vigorous PA (MVPA) across the week as compared to controls with (OC) and without obesity (LC) [51]. In a secondary analysis of these data, Creasy et al. found that weight loss maintainers were most active in the morning and averaged 3278 ± 3004 steps and 25.1 ± 23.1 min of MVPA (~27% of daily totals) within 3 hours of waking, significantly higher levels than controls (Figure 2) [49]. In a separate study, Schumacher et al. also found that weight loss maintainers self-reported engaging in high amounts of MVPA [46]. In that study, participants who were classified as temporally consistent exercisers (i.e., regularly exercised at the same time of day) engaged in more MVPA (350 min/wk) compared to temporally inconsistent exercisers (285 min/wk). However, the time of day that activity was performed was not associated with amount of MVPA. Thus, in the context of chronic exercise and achieving high levels of PA, consistency of exercise may be more important than time of day. Engaging in PA at the same time of day may help with the habit formation necessary for achieving high levels of MVPA. However, cross-sectional data for which this premise is based on should be interpreted with caution. Prospective research is needed to understand the physiological and behavioral mechanisms through which exercise timing and consistency of PA influences body weight regulation.

Figure 2:

24-hour patterns of stepping in weight loss maintainers (WLM), lean control (LC) and control participants with overweight/obesity (OC)

In summary, exercise at any time of the day is beneficial for weight management and several other health outcomes. Some studies suggest that morning exercise may promote superior weight management benefits as compared to evening exercise [42, 45, 49]. The effects appear minimal in one short-term randomized study of 6 weeks [42]; however, in a 10-month non-randomized study those who performed most of their exercise in the morning achieved 3 times greater weight loss than those who performed most of their exercise in the evening [45]. The data are equivocal with two non-randomized short-term studies finding no differences in weight loss, and superior fat loss in evening exercise compared to morning exercise [44, 48]. However, these collective data come from small studies, secondary analyses, and observational studies. Adequately powered randomized trials are needed to confirm these preliminary findings. A major area of interest is understanding the mechanisms by which exercise timing may generate differences in weight loss. The remainder of the review will focus on 3 major areas which may contribute to the impact of exercise timing on body weight regulation: 1) EI, 2) EE and substrate metabolism 3) and sleep.

EFFECT OF EXERCISE TIMING ON ENERGY INTAKE

In general, acute bouts of exercise have been shown to alter hormonal drives of EI leading to reductions in hunger and short-term EI; however, this effect is not observed in all individuals and may be influenced by sex, adiposity, and the exercise stimulus [52–54]. Some longer-term studies examining the effect of exercise training on EI show that exercise leads to a compensatory increase in EI [18, 55–58]; however, other studies have found minimal increases in EI in response to exercise training [16, 37, 59]. In these studies, there is substantial inter-individual variability in the EI response to exercise. Time of day of exercise may be an underappreciated factor contributing to this variability. Well-controlled studies have found that hunger and appetite hormones such as leptin, ghrelin, and peptide YY exhibit circadian oscillations [60–62]. Because appetite related hormones exhibit circadian rhythmicity and are influenced by exercise, morning versus evening exercise may differentially impact subsequent EI. However, current evidence from well-controlled studies on time of day of exercise and its effect on hunger, drivers of appetite, and total EI is limited.

Comparison of Morning vs. Evening Exercise – Acute Bouts

In a randomized cross-over study by Alizadeh et al. [63], 50 women with overweight engaged in a 30 minute bout of exercise at the ventilatory threshold in the morning (08:00–10:00) and the afternoon (14:00–16:00). Visual analog scales were used to measure hunger, satiety, and food cravings before and 15 minutes following exercise. Subsequent EI was measured over the 24 hours following each exercise session using 24-hour food records. Although EI was not different, satiety was higher following morning exercise compared to evening exercise suggesting that participants felt fuller following morning exercise. In a separate study by Maraki et al. [64], 12 young, normal weight females performed either 1 hour of exercise or 1 hour of seated rest in the morning (08:15–09:15) or evening (19:15–20:15). Both morning and evening exercise induced a similar 24-hour energy deficit, increased hunger and prospective food consumption, and decreased satiety and fullness compared to control conditions, but there were no differences between conditions. Finally, Larsen et al. [65] compared appetite and appetitive hormones in response to acute morning (06:00–07:00), afternoon (14:00–16:00), and evening (19:00–20:00) exercise. Although ghrelin was higher 30 minutes after afternoon exercise compared to morning and evening exercise, subjective appetite responses after exercise were not different between groups. Additionally, self-reported 24-hour EI following the exercise bout was not different between all three conditions. Thus, timing of an acute exercise bout does not seem to impact subsequent EI.

Comparison of Morning vs. Evening Exercise – Exercise Training

Few studies have investigated the impact of exercise timing on EI and appetite responses in the context of chronic exercise training. Two studies which demonstrated superior weight loss with morning exercise compared to evening exercise have also documented a potential time of day effect on EI and appetite [42, 45]. In the study by Willis et al., EI was quantified using digital photography in a school cafeteria setting assessed over 7 days pre- and post-intervention. Food consumed outside of the cafeteria was assessed using multiple-pass recalls. Despite significant differences in weight loss outcomes between morning and evening exercisers, there were no significant within or between group differences in energy intake over 10 months (morning exercisers: −63±444 kcal/d; evening exercisers +121±484 kcal/d) [45]. However, this study was not powered to detect differences in EI and these small observed differences in EI may have influenced weight loss. Alizadeh et al. found that morning exercise tended to reduce 24-hour self-reported EI by 362 kcal/d compared to evening exercisers who reported a change in EI of −28 kcal/d (p=0.06 for group difference) [42]. However, neither group reported any consistent changes in hunger, satiety, and other eating behaviors across the exercise intervention. Finally, Di Blasio et al. showed no differences in change of total EI or hunger/satiety following 3 months of morning and evening exercise, but there were significant changes in the timing of EI [44]. Evening exercise increased the proportion of EI consumed in the morning and decreased the proportion of EI consumed in the evening, while morning exercise did not alter timing of food intake. This study is one of the first to demonstrate that the timing of exercise may naturally shift timing of food intake.

A major and perhaps under recognized challenge in understanding the impact of exercise timing on EI is related to the time of the last EI event. The duration of fasting prior to a morning exercise bout compared to an evening exercise bout can be very different. Bachman et al. found that individuals who performed fasted morning exercise vs. non-fasted morning exercise reported consuming less EI over the next 24h [66]. These data demonstrate that fasting prior to exercise may affect EI, thus it is important to consider the duration of fasting when examining the effect of exercise timing on appetite and EI responses. Given the logistical challenges of controlling fasting windows prior to exercise at varying times of the day, this may be a complicated area of study. Nonetheless, more research is needed in this area, as it is difficult to draw conclusions from existing literature. Most of the studies to date have relied on self-reported measures of EI, appetite/hunger, and eating behavior. Future controlled studies are needed to understand whether the timing of exercise alters objectively measured food intake over periods longer than 24 hours, appetite-related hormones, and neuronal responses to food cues.

EFFECT OF EXERCISE TIMING ON ENERGY EXPENDITURE AND SUBSTRATE OXIDATION

Exercise likely has differential metabolic effects depending on the time of day it is performed; however, this is an understudied area that needs further exploration. Several studies using circadian protocols (e.g., constant routine, forced desynchrony, simulated shift work) have shown that resting metabolic rate [60, 67–69], thermic effect of food [70], and substrate oxidation [60, 67–69, 71] vary over 24h; however, the rhythmicity and timing of peaks and nadirs have varied by study. Despite the inconsistency in the rhythmicity of these data across studies, it is clear that whole-body metabolism changes across the day in humans. While research to date has focused on resting metabolism, it is likely that exercise metabolism (i.e., substrate utilization and EE) also varies across the day. It is possible that exercise timing could alter EE during an exercise bout or over the subsequent 24 hours, which would have implications for weight management. In addition to physiological effects, the timing of exercise may alter diurnal behaviors for the rest of the day. For example, morning exercise may invigorate a person to reduce sedentary and sitting behaviors and increase non-exercise PA the rest of the day. In contrast, it is possible that morning exercise results in increased fatigue the rest of day and results in increased sedentary behavior and decreased non-exercise PA. Below, we outline the existing evidence on physiological and behavioral mechanisms through which the timing of exercise may affect whole body metabolism.

Two recent studies in mice found that exercise time of day altered molecular metabolic pathways, fuel utilization, exercise capacity and whole-body EE [72, 73]. Interestingly, Ezagouri et al. also found that in 10 adults, evening exercise (performed at 18:00) resulted in lower oxygen consumption, higher respiratory exchange ratio (e.g., more carbohydrate metabolism), and lower perceived exertion compared to morning exercise (performed at 08:00) at the same workload over 60 minutes [73]. The authors concluded that evening exercise is more metabolically efficient than morning exercise. Whether differences in substrate utilization and metabolic efficiency translate to differences in exercise EE, weight loss or other clinically relevant measures of health is unclear and worthy of future study.

Few studies have examined the effect of the exercise time of day on PA behaviors. Bond et al. [74] found that in a population of bariatric surgery patients regular morning exercisers (defined as individuals who performed their longest bout of MVPA between 04:00–12:00) had higher levels of adherence to the exercise intervention and higher increases in total PA compared to those who performed most of their exercise after 12:00. The authors concluded that prescribing morning PA may be an effective strategy to increase MVPA in bariatric surgery patients. In the study by Willis et al., compared to evening exercisers, morning exercisers engaged in higher amounts of non-exercise PA and lower amounts of sedentary behavior compared to evening exercisers [45]. This was associated with higher levels of non-exercise EE. Although these differences were not significant, if they were consistent over many months, they could significantly affect bodyweight. No studies to date have been adequately powered to study the effect of exercise time of day on non-exercise PA and/or changes in EE.

EFFECT OF EXERCISE TIMING ON SLEEP

Current evidence suggests that acute bouts of exercise [75, 76] and chronic exercise training [75–77] improve subjective sleep quality and objectively derived measures of sleep. Because both sleep quality and sleep duration have been linked to bodyweight regulation [78–81], it is possible that sleep is a mediator in the relationship between exercise and weight loss. Below we review existing literature on whether the effect of acute and chronic exercise on parameters of sleep is dependent on time of day.

Comparison of Morning vs. Evening Exercise – Acute Bouts

Studies directly comparing aerobic exercise performed at different times of day on measures of sleep (e.g., sleep architecture and sleep quality) are limited. Research on the effect of exercise timing on measures of sleep have primarily focused on the potential sleep disrupting effects of evening exercise. Evening exercise was originally believed to be detrimental to sleep quality, however a recent meta-analysis by Stutz et al. [82] reported that evening exercise improved sleep quality compared to no exercise. A separate meta-analysis by Kredlow et al. [76] showed that time of day of moderated the beneficial effects of exercise on wake after sleep onset (WASO) and stage 1 sleep but not total sleep, sleep efficiency, sleep onset latency (SOL), and slow wave sleep. In a few small studies, morning exercise has also shown favorable effects on indices of sleep quality. Fairbrother et al. showed that compared to afternoon exercise (13:00), morning exercise (07:00) reduced sleep latency and increased time slow wave sleep in middle-aged adults with prehypertension [83]. Similarly, Morita et al. [84] showed that a group of older adults with insomnia reduced awakenings during the night after an acute bout of morning exercise (09:30) compared to evening exercise (17:30). While some data support a benefit of exercise regardless of time of day, the limited number of head-to-head comparisons of exercise times across the day make it difficult to discern the best time of day for exercise to improve sleep quality.

One factor that may be important in optimizing exercise for improvements in sleep quality is the impact of chronotype. Chronotype describes an individual’s natural propensity for sleep and wake timing [85]. A recent observational study of sleeping patterns after acute exercise in 909 college students found that chronotype was a moderator of the relationship between exercise timing and bed timing [86]. Later exercise times were associated with later bedtimes in both morning and evening chronotypes; however, the effect was more pronounced in morning chronotypes. These data demonstrate that the impact of exercise timing on sleep may be dependent on an individual’s natural propensity for wake and sleep. Future studies should include assessments of chronotype when investigating the optimal timing of exercise for sleep.

Thomas et al. studied the effect of exercise timing on an individual’s sleep and circadian rhythm (dim light melatonin onset, DLMO) [87]. Participants were randomized to perform 5 days of supervised morning exercise (10 hours after DLMO) or evening exercise (20 hours after DLMO) for 30 minutes per day at 70% VO₂ peak. While there were no differences in sleep duration or fragmentation (assessed by actigraphy), there was a shift in the melatonin rhythm with exercise. Regardless of chronotype, morning exercise resulted in a phase advance of the melatonin rhythm (i.e., DLMO occurred at an earlier time); however, the impact of evening exercise was dependent on chronotype. Evening exercise resulted in a similar phase advance of DLMO in evening chronotypes, but a phase delay (i.e., DLMO occurred at a later time) in morning chronotypes. These data highlight emerging evidence that exercise may act as a zeitgeber (time cue) for the central circadian system. Youngstedt et al. performed a rigorous circadian study finding that exercise performed in the morning (07:00) and early afternoon (13:00 and 14:00) led to a phase advance in melatonin rhythms, whereas exercise in the evening (19:00) and at night (22:00) resulted in a phase delay [88]. Exercise at 02:00 and 16:00 had a minimal impact on the melatonin rhythm. Collectively, these data indicate that there are times of the day that the circadian system may be more sensitive to exercise. Ultimately, alterations to the circadian system will likely have downstream effects on other metabolic processes and may be a mechanism responsible for differential impacts of timed exercise. Future studies should leverage our current understanding of exercise on central circadian rhythms to guide future explorations of the therapeutic potential of exercise timing for sleep, metabolism, and weight management.

Comparison of Morning vs. Evening Exercise – Exercise Training

In the context of exercise training, much less is known regarding the impact of exercise timing on sleep. Overall, exercise training studies show improvements in sleep quality [89–94]; however, most studies have allowed participants to self-select the time of exercise. Very few studies have randomized participants to perform exercise at different times of day. A short-term training study by Benloucif et al. showed similar improvements in subjective sleep quality with morning and evening exercise training for 2 weeks in older adults [95]. Küüsmaa-Schildt et al. randomized 70 young men to perform either morning or evening exercise (~2 days per week of combined endurance and strength) training for 24 weeks. Both morning and evening exercise groups experienced small improvements in self reported sleep quality and duration [96]. Finally, Seol et al. randomized 60 older adults to eight weeks of either a morning or evening exercise training (30 min/day of low intensity stepping). Both exercise groups showed similar improvements in objectively derived measures of sleep quality including sleep efficiency and wakenings after sleep onset [97]. Together, these data suggest that the sleep quality benefits of chronic exercise training may be more related to consistently performing exercise rather than the time of the exercise bout. Notably, the majority of sleep and exercise research has been conducted in healthy young men, many of whom are highly trained athletes [98–104]. Therefore, a major gap in the current literature is a lack of studies in populations that are translatable to the overall population. Future studies with more diversity of sex and gender, race/ethnicity, BMI, and chronotype are necessary to conclusively determine the impact of exercise timing on sleep within the context of exercise training.

RESEARCH GAPS AND FUTURE DIRECTIONS

Does exercise timing affect weight loss success?

There is a mounting evidence demonstrating that the time at which behaviors are performed confer differential health benefits, particularly related to weight management. As highlighted by this review, most studies that have investigated the impact of exercise timing on weight loss are comprised of small pilot studies, secondary analyses, and observational studies. It is difficult to confer a definitive conclusion from these studies. Adequately powered, prospective, randomized trials in generalizable cohorts of both men and women with overweight and obesity are needed to confirm these preliminary findings. In particular, studies are needed that are longer in duration, prescribe exercise based on EE at a dose that will elicit weight loss, and ensure high adherence to exercise at a dose that will elicit weight loss. Future studies are also needed to examine if sex, gender, race/ethnicity, and baseline BMI moderate the relationship between time of day of exercise and weight loss. By identifying individuals who benefit most from exercise at different times of day, researchers, clinicians, and public health officials can create more tailored treatment paradigms which will result in better overall treatment success.

How does the timing of exercise affect energy balance?

There is lack of understanding of both the behavioral and physiological mechanisms through which time of day of exercise may differentially affect body weight. To better understand the mechanisms by which exercise timing may confer differential effects on energy balance and body weight, both acute and chronic training studies are needed. Acute studies are needed to understand how timing of exercise bouts may affect factors such as eating behavior, appetite-related hormones, meal-timing, overall EI, substrate utilization, exercise EE, overall daily EE and other factors related to energy balance. Training studies are needed to understand how consistent training at the same time of day may elicit differential effects on these parameters as compared to acute exercise. For example, consistent exercise training at the same time of day may lead to shifts in meal timing, sleep timing, and other factors that would not be detectable in short term studies. Moreover, studies are needed which anchor exercise timing to sleep/wake rather than clock time. While using clock time is a practical approach, using habitual sleep/wake time to anchor exercise may help to reduce variability in outcomes of interest.

Does the timing of exercise affect sleep?

Having better sleep quality may affect both sides of the energy balance equation by increasing PA, decreasing sedentary behaviors during the day [105] and improving dietary behaviors. While exercise is believed to be important for sleep, the importance of exercise timing for sleep remains unknown. Both acute, in-depth sleep studies and exercise training studies are needed to understand if there is an optimal time of day to exercise for improving sleep. Moreover, future studies are needed with more diverse populations, including women, untrained individuals, adults across the age and BMI spectrum, and adults with various health conditions. Specifically, future studies with individuals who have reported sleep disturbances are needed to understand if timed exercise is a potential therapeutic target.

CONCLUSIONS

Exercise at any time of day is important for overall health. Exercise and PA are critically important to treat and prevent overweight and obesity; however, there is considerable individual variability in the weight loss response to exercise. The timing of exercise may be a contributing factor to this observed variability in weight loss. As illustrated in Figure 1, the timing of exercise may have a differential impact on EE, EI and sleep, with all these factors ultimately affecting body weight regulation. Studies comparing morning versus evening exercise and their effect on body weight regulation are inconclusive. Ultimately, large-scale clinical and translational studies are needed to if altering time of day of exercise can lead to clinically meaningful differences in weight loss and other health outcomes.