<b>PROTOCOL</b> : Examining the best time of day for exercise: A systematic review and network meta‐analysis

Global health is influenced by factors such as ageing population, rapid urbanisation, and increased transportation by car and plane, all of which result in unhealthy environments and behaviour. As a result, the growing prevalence of noncommunicable diseases (NCDs) and their risk factors has become a global issue (Zhang et al., 2017). According to the WHO World Health Statistics 2018 (WHO, 2018), In 2016, an estimated 41 million deaths occurred due to NCDs, accounting for 71% of the overall total of 57 million deaths. The majority of such deaths were caused by the four main NCDs, including: cardiovascular disease (17.9 million deaths; accounting for 44% of all NCD deaths); cancer (9.0 million deaths; 22%); chronic respiratory disease (3.8 million deaths; 9%); and diabetes (1.6 million deaths; 4%). There are many key risk factors of NCDs such as tobacco use, air pollution, unhealthy diet, physical inactivity and harmful use of alcohol, one of the main risk factors is physical inactivity (WHO 2018). Physical inactivity has been identified as the fourth leading risk factor for global mortality (6% of deaths globally), with an estimated 20%–30% increased risk of death compared with those who are physically active. This follows high blood pressure (13%), tobacco use (9%) and high blood glucose (6%). Overweight and obesity are responsible for 5% of global mortality (WHO, 2013). It has been shown that participation in regular physical activity reduces the risk of coronary heart disease and stroke, diabetes, hypertension, colon cancer, breast cancer and depression. Additionally, physical activity is a key determinant of energy expenditure, and thus is fundamental to energy balance and weight control (WHO, 2002, 2005, 2007, 2008, 2010). Physical activity is defifined as bodily movement produced by skeletal muscle contraction that requires energy expenditure above basal levels. It includes activities related to activities of daily life, such as housekeeping, yardwork, occupational-related, leisure-related and transportation. Exercise typically is differentiated from physical activity in that it is typically planned, repetitive, and structured with the main objective of improving health and fifitness. Physical fifitness is a state of good health and strength achieved through physical activity and exercise (Fletcher et al., 2018). The main types of physical exercise include aerobic exercise, resistance training (anaerobic exercise), flexibility and balance. Aerobic exercise increases the uptake of oxygen in the larger muscle groups and has a beneficial effect on cardiovascular homoeostasis. Resistance training mainly affects muscle strength and mass whereas flexibility and balance exercise improve the range of motion necessary for daily activity and diminish the risk of falls (Nehrlich, 2006; Nelson et al., 2007). Being physically active and exercise is one of the most important actions individuals of all ages can engage in to improve their health. The evidence reviewed by the Physical Activity Guidelines Advisory Committee 4 for the newly released Physical Activity Guidelines for Americans, 2nd edition 5 (PAG) is clear—physical activity fosters normal growth and development and can make people feel better, function better, sleep better, improve wellbeing and reduce the risk of many chronic diseases (FUüzeki et al., 2017; Hills et al., 2015; Nehrlich, 2006; Warburton Darren & Bredin Shannon, 2017). However, many factors may influence the benefits of doing regular exercise, such as the exercise type, duration, hormone adaptation, and timing of exercise (Seo et al., 2013). To maximise the benefits of exercise, more tailored exercise prescriptions are required. Abundant scientific evidence has demonstrated that time of day is one of the important factors that affects the outcomes related to exercise. However, there remains much debate when to exercise and when to do what type of exercise. Is exercising in the evening, mid-day or in the morning is better for human health? The existence of a time-of-day effect on human performance is now well established. However, the optimal timing of exercise for health have not been fully established (Kraemer et al., 2001). Numerous physiological phenomena in the human body, such as sleep–wake cycles, hormonal and nervous activity, and body temperature, exhibit rhythmic changes over the course of 24 h (Bass & Takahashi, 2010; Yu & Shibata, 2013). It is already known that hormone concentrations exhibit circadian rhythmicity and so vary throughout the day (Kraemer et al., 2001) along with body temperature (Bailey & Heitkemper, 2001) and strength performance (Sedliak et al., 2009). With the different hormone concentrations, motor function (such as strength, whole-body flexibility, simple reaction time and short-term power output) display a time-of-day effect characterised by a late afternoon acrophase (~18:00 h) and an early morning bathyphase (~06:00 h; see reviews by Atkinson & Reilly, 1996, Reilly et al., 2000). More specific changes of body indicators in 24 h are shown in Figure 1, we can see in the morning (06:00–10:00 h), afternoon (16:00–18:00 h), and evening(19:00–21:00 h), our body is the most energetic and suitable for exercise, which needs to take a lot of energy and strength. However, research on optimising exercise timing is sparse. In other words, which is the best time of the day (morning, afternoon and evening) to trigger an optimal training response remains unclear (Gabriel & Zierath, 2019). The Physical Activity Guidelines for Americans, 2nd edition (Piercy et al., 2018), provides information and guidance on the types and amounts of physical activity that provide substantial health benefits. But it has no recommendation on the best time of a day to exercise. There are several RCTs (Brito et al., 2018; Drust et al., 2005; Gabriel & Zierath, 2019; Souissi et al., 2002) focusing on the timing of exercise. One RCT with 10,086 participants which indicated that body weight, body mass index, abdominal skin fold thickness and abdominal circumference decreased significantly more with morning exercise compared to evening exercise over a 6 week period (Gabriel & Zierath, 2019). Souissi et al (Souissi et al., 2002) suggested that several weeks of repeated strength training performed in the morning hours may reduce the typical diurnal pattern by increasing maximum strength more in the morning than at other times of day. However, the results about different timing for exercise are not always consistent. For example, Drust et al. (2005) reported that maximum voluntary strength of untrained men typically exhibits a diurnal pattern, with low morning values and peak values in the afternoon. And it's worth noting that the effect of time on exercise effects may vary by age, as people of different ages have different physical qualities. Knowing the right time to do exercise could be valuable for people to improve the quality of life, but the conflicting results from current evidence make it difficult to draw a conclusion to this question. Therefore, it is necessary to systematically gather and synthesise published and unpublished evidence to examine the best time of day for different types exercise. The primary objective of the review is to synthesise evidence on the effectiveness of exercise in different times of a day. The review aims to answer the following question: Which is the best time of day for different types of exercise—morning, afternoon or evening—to obtain the highest benefits such as physical and mental health, weight loss and so on? The proposed review will include following three types of studies: (1) randomised controlled trials in which participants are randomly assigned to an experimental or control group, (2) quasi-randomised controlled trials in which participants are allocated by means such as alternate allocation, person's birth date, the date of the week or month, or alphabetical order, and (3) nonrandomised controlled trials in which participants are nonrandomly assigned to an experimental or control group. In these designs, studies compared exercise during different times of day, or studies that compared exercise during a specific time of day with no exercise, for example, a study compared exercise in the morning to exercise in the afternoon; or studies that compared exercise in the morning with no exercise, the experimental group refers to participants who exercise in the morning, afternoon, or evening to improve their health, and the control group refers to those in no-training. We will also include studies comparing different types of exercise at the same or different specific time of day. The review will include all human populations including any age, sex and health status. Studies will be included if their interventions meet the following criteria: (1) included exercise, that is physical activity comprising planned, structured and repetitive body movements, which are undertaken to improve one or more components of physical fitness according to the American College of Sports Medicine (Ferguson, 2014); (2) compared exercise during different times of day, or studies that compared exercise during a specific time of day versus no exercise. The primary outcome measures include physical health, mental health, general health/ill health and quality of life indicators measured using validated instruments (some possible examples are listed in Table 1). General Health Questionnaire (GHQ-12) RAND Mental Health Inventory (MHI) Hospital Anxiety and Depression Scale (HADS) Warwick-Edinburgh Mental Well-being Scale EORTC (European Organisation for Research and Treatment of Cancer) Quality of Life Questionnaire Functional Limitations Profile (FLP) Short Form (SF-36) EuroQol (EQ-5D) We will extract outcomes relating to social wellbeing,specifically work-life balance, but only when reported alongside primary outcomes and when work-life balance was measured using a validated instrument. We will also extract some anthropometric indices, such as heart rate, calorie intake, macronutrient consumption, weight loss and so on, shown in Table 2. Social/domestic disruption/interference Family-to-work conflict Time spent with friends/family Time spent on domestic chores/hobbies Studies with any length of follow up will be included in this review. In order to synthesise data from studies with different lengths of follow-up, groups will be defined and analysed separately (eg., 0–6, 6–12 and >12 weeks) and will be pooled where there are no significant differences. Studies that report participants in any setting will be included. We will search electronic databases, grey literature sources and hand search journals to identify all potentially eligible studies. There will be no restrictions placed on document language or publication status. We will also contact leading authors and experts in the field of exercise and health for additional studies via email. The bibliographies of relevant reviews and included studies will be hand searched to identify additional references for review. The search strategy of Medline is as follows: "Exercise"[Mesh]) OR Exercise*[Title/Abstract] OR Physical Activit*[Title/Abstract]) OR Training[Title/Abstract] AND "Time"[Mesh] OR Time[Title/Abstract] OR timing[Title/Abstract] OR Morning[Title/Abstract] OR Afternoon[Title/Abstract] OR Evening[Title/Abstract] OR schedule[Title/Abstract] OR night[Title/Abstract] OR noon[Title/Abstract] Medline Embase Web of Science Cochrane Central Register of Controlled Trials (CENTRAL) Chinese Biomedical Literature Database (CBM) China Network Knowledge Information (CNKI) Wanfang Database Social Care Online: https://www.scie-socialcareonline.org.uk/. AgeLine: http://www.aarp.org/research/ageline/. Global Health: http://www.cabi.org/datapage.asp?iDocID=169. Database of Promoting Health Effectiveness Reviews (DoPHER): http://eppi.ioe.ac.uk/webdatabases/Intro.aspx?ID=2. We will search Google Scholar, OpenSIGLE (http://opensigle.inist.fr/), Opengrey (http://www.opengrey.eu/), Clinicaltrials.gov (https://www.clinicaltrials.gov/), World Health Organization International Clinical Trials Registry Platform (https://www.who.int/ictrp/en), Social Science Research Network (https://www.ssrn.com/index.cfm/en/) and major sports science institutions (https://www.lboro.ac.uk/departments/ssehs/; https://www.acsm.org/) for grey literature. Copies of relevant documents will be made and we will record the exact URL and date of access for each relevant document. In addition to check the reference lists of included studies and reference lists of relevant reviews, we will also hand search following relevant journals: Medicine and science in sports and exercise Psychology of sport and exercise Research quarterly for exercise and sport Scimago Institutions Rankings (https://www.scimagojr.com/journalrank.php?Category=3699), which included many journals of sports, such as British Journal of Sports Medicine, American Journal of Sports Medicine and so on. Endnote X9 software will be used to manage retrieved bibliographies. EPPI-Reviewer 4 will be used to screen retrieved bibliographies and extract data. After the removal of duplicate results, two reviewers will first independently screen titles and abstracts to exclude studies that are clearly irrelevant. If studies are considered eligible by at least one assistant or there is insufficient information in the title and abstract to judge eligibility, will be retrieved in full text. The selected review author pair will collect full-text trial publications, and independently screen the full-texts and identify trials for inclusion, any disagreement of eligibility will be resolved by a third party from the review authors. Exclusion reasons for studies that otherwise might be expected to be eligible will be documented and listed in an appendix. We will record the selection process in sufficient detail to complete a PRISMA flow diagram and ‘Characteristics of excluded studies’ table for studies excluded on full text. Two reviewers, working in pairs, will independently extract data using data extraction forms designed for the purpose, and we will pilot the form against sample studies before finalising. If there is disagreement, the authors will discuss the reasoning behind their assessment. If an agreement is not reached between the two authors, YL will serve as arbitrator. The following information will be extracted from each included study: Document description (e.g., first author, year of publication, country) Methodological issues (e.g., design, randomisation, blinding, selective reporting) Participant information (e.g., age, gender, health status, muscle mass, fat mass, body fat) Intervention (e.g., exercise time, exercise type, exercise duration) Outcomes (e.g., wellbeing, sleep quality, weight loss, heart rate) The length of follow up Two authors will independently assess the risk of bias in each included study. Discrepancies will be discussed with a third author until consensus is achieved. Randomised controlled trials and quasi-randomised controlled trial will be assessed using the tool recommended by the Cochrane Handbook Version 5.1.0 (Higgins & Green, 2011), We will assess the following domains: risk of bias, allocation sequence generation, allocation concealment, blinding of outcome assessors, incomplete outcome data, selective outcome reporting and other potential sources of bias (i.e., the length of training and confounding variables). RCTs will not be assessed for blinding of participants as the participatory nature of interventions (exercise in different time) makes blinding impossible. When the risk of bias of all seven terms is defined as “low risk of bias”, the trial will be defined as the overall “low risk of bias”. At the same time, when one or more of the seven bias components are classified as high risk, the trial will be graded as “High risk of bias”. In other cases, the trial will be graded “Unclear risk”. Non-RCTs will be evaluated according to ROBINS-I In Studies of et al., which is a tool for risk of bias in of the effectiveness or of interventions from studies that not use to or of to such as studies and studies in which groups are allocated during the course of and quasi-randomised studies in which the of allocation falls of full ROBINS-I includes seven the first two confounding and selection of participants the issues before the of the interventions that are to be compared The third of the interventions The other four issues the of due to from data, of outcomes and selection of the reported For such as of with chronic we will the and using the When the for outcomes of we will an in of with which will be considered to a potential effect and a of will be to studies that reported For such as quality of life and sleep quality, the included studies may use for the same with will be to synthesise the that are If the of the same outcome are for different with their will be to synthesise the If studies not report will be from and using proposed by the Cochrane Handbook (Higgins & Green, If a of OR and was reported for one we will the to using the in Figure reported by et al. and For trials with more than two we will the group two or more groups with sample and include two or more et al., For example, in a trial with three and where the of are the of outcome in versus and versus we will the sample of by two and it to the and the of outcome We will use this group in the against and et al., We will identify incomplete outcome data during data the trial report that outcome data were but not we will contact the author to data. a trial has been and a relevant outcome was in the trial but no results we will contact authors and to trial We will the and of the included studies by comparing gender, and the health interventions type of exercise, the time of follow of exercise), using information reported in the of included In we will examine using will also be assessed through and and was considered of and high (Higgins & When the is than we will and to possible reasons for In can be considered an additional of which can of It can when there is a between a and of effect et al., will use the to the effect and for the and for the of between and et al., We will these on the primary If or more studies are included in an we will use to assess publication bias at the study outcome reporting will be assessed as of the of the risk of bias in the included studies. at least two studies are we will for all outcomes and with data from two or more We will assess by comparing the and will not groups in a We will in using the and effects We will all randomised participants according to the with the that participants who were to not to We will in and the of the effect measures for each (i.e., and for each and the OR with Network a will be used to and evidence from all and the on a of the software will be used to process the The function of of be used to to and the of different times of exercise. The will be used to different times and to the between will be reported with and a will be considered effects will be used to pooled and it the between use for For primary we will display the of interventions by the the et al., which the that one is better than other We will a for with values for in the and values for in the et al., 2013). Studies in the right hand will be considered to have the best balance of and If we we will and do if we in we will use will be used as the overall of This reduces the of type by of effects by on the in effect We will systematically examine all possible as which may influence the effects of exercise. in participants (e.g., gender, age, The time of follow up (e.g., 0–6, 6–12 and >12 The health (e.g., cancer chronic disease The type of exercise (e.g., aerobic activity, balance We to on trials classified as high quality versus trials classified as low or RCTs versus We will assess the of any study that has a effect on the results of the We will use of and et al., to assess the of evidence with specific outcomes and a of The is used to assess the quality of a body of evidence on the to which one can be that an of effect or the of the quality of evidence risk of bias, publication bias and other bias et al., We do not to include We to the of the and reviewers of the Social group, and in who a lot in this And also for the by the of the Social Science of on the International and Chinese of Social and review and and Information and is the of the other authors have no of date for of the and will be responsible for the 5 references

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.