Physical activity and sedentary behaviour in the Middle East and North Africa: An overview of systematic reviews and meta-analysis

To support the global strategy to reduce risk factors for obesity, we synthesized the evidence on physical activity (PA) and sedentary behaviour in the Middle East and North Africa (MENA) region. Our systematic overview included seven systematic reviews reporting 229 primary studies. The meta-analysis included 125 prevalence measures from 20 MENA countries. After 2000, 50.8% of adults (ranging from 13.2% in Sudan to 94.9% in Jordan) and 25.6% of youth (ranging from 8.3% in Egypt to 51.0% in Lebanon) were sufficiently active. Limited data on PA behaviours is available for MENA countries, with the exception of Gulf Cooperation Council countries. The meta-regression identified gender and geographical coverage among youth, and the PA measurement as predictors of PA prevalence for both adults and youth. Our analysis suggests a significant PA prevalence increase among adults over the last two decades. The inconsistency in sedentary behaviour measurement is related to the absence of standardized guidelines for its quantification and interpretation. The global epidemic of insufficient PA is prevalent in MENA. Lower PA participation among youth and specifically females should be addressed by focused lifestyle interventions. The recognition of sedentary behaviour as a public health issue in the region remains unclear. Additional data on PA behaviours is needed from low- and middle-income countries in the region.

Non-communicable diseases (NCDs) kill 41 million people worldwide each year -equivalent to 71% of all deaths 1 . The Middle East and North Africa (MENA) region has one of the highest rates of NCDs in the world. In 2017, the region reported the second highest prevalence of diabetes in the world (10.8%) 2 and is recording a rapid increase in obesity 3,4 . Insufficient physical activity (PA) and sedentary behaviour are key risk factors for obesity and other NCDs [5][6][7][8][9][10][11][12][13][14][15] leading to premature mortality 10,11,[16][17][18] . It has been suggested that PA has the potential to effectively control and reduce the burden of obesity during the various phases of human development 19 . Regular PA can also improve self-esteem, cognitive performance, and academic achievement in young people 7,20,21 and is positively related to cardiorespiratory and metabolic health 6 . Recently, sedentary behaviour has received global attention as prolonged sedentary time is associated with an increased risk of chronic disease and an increase in all-cause mortality, regardless of individuals meeting the recommended levels of PA 13,15 .
The World Health Organization (WHO) and the Global Observatory for Physical Activity (GoPA) are targeting a relative reduction of 10% in the global prevalence of physical inactivity among adults by 2025 22,23 . Currently, one of the most pressing needs to improve population health is to develop appropriate policies and implement interventions to address the global pandemic of physical inactivity 24,25 . However, to support this action, country-level evidence on PA behaviour in various population groups is essential. Both regional-and country-level data contribute to the continuous surveillance of PA participation and are essential to track progress towards the global PA target.
In the MENA region, the proportion of the population not engaging in sufficient levels of PA (as per the recommendation of standardized international guidelines) remains unclear. Recent reports indicate physical inactivity prevalence measures of 32.8% for adults across the MENA and Central Asia regions 26

and 78.4% in boys
Institute for Population Health, Weill Cornell Medicine-Qatar, Doha, Qatar. ✉ e-mail: kac2047@qatar-med.cornell.edu open Search strategy and selection criteria. A broad search strategy was developed to systematically identify any type of review on all health issues in any MENA country. Search terms related to countries' names, MENA populations' names, and MENA sub-regions' names, such as North Africa, East Africa, and the Middle East, were used. No restrictions to a specific health condition or intervention and language of publication were applied at this stage. The full search strategy with search criteria is available in the published overview protocol 27 . Two independent reviewers (AA and HA) systematically searched the Medical Literature Analysis and Retrieval System Online (MEDLINE) through the search engine PubMed. We included publications since 2008 -the publication year of the first version of the Cochrane Handbook for Systematic Reviews of Interventions 30 up to February 21, 2017. Additionally, we also searched grey and non-grey literature sources with no date or language restriction including Google Scholar, Epistemonikos, ProQuest, OpenGrey, Bioline International, E-Marefa, ALMANHAL platform, governmental websites of all MENA countries, and the WHO website for systematic reviews relevant to our topic. The literature search was then updated to identify recent SRs published up to November 2019.
A systematic review (SR) was defined as a literature review that has explicitly used a systematic literature search of at least one electronic database to identify all studies that meet pre-defined eligibility criteria along with a study selection 30 . Reviews not reporting a systematic methodology, such as narrative reviews, were excluded.
Based on the relevance of grey literature in the region when studying population health outcomes 27,31 , we included MENA countries where Arabic, English, French, and/or Urdu are the primary official languages and/ or the medium of instruction in the colleges/universities. These languages are also those spoken by the authors of this overview (see the overview's protocol 27 ). The 20 MENA countries included are Algeria, Bahrain, Djibouti, Egypt, Iraq, Jordan, Kuwait, Lebanon, Libya, Morocco, Oman, Pakistan, Palestine, Qatar, Saudi Arabia, Sudan, Syria, Tunisia, the United Arab Emirates (UAE), and Yemen. Data screening. Records were downloaded into Endnote (version X8.2), and duplicates were removed.
Using Rayyan software 32 , two independent reviewers (AA and HA) conducted the multi-stage screening following a standard process. Discrepancies in the inclusion of SRs were resolved through discussion, with the involvement of a third reviewer (AA, HA, and KC).
Retrieved SRs were then categorized based on the reported population health outcome. For the purpose of this overview, we included any SR reporting measurable PA-related outcomes including physical activity or inactivity prevalence and/or sedentary behaviour pertaining to MENA population of any age group residing in a MENA country.
Data extraction. Data extraction was conducted by SC1 and checked for accuracy by KC. Extracted data included characteristics of the included SRs as well as the primary studies and were matched to PICOTS items (Population, Outcomes, Time of the study and Setting; Control and Intervention were not applicable to our overview question). From each included SR, the following characteristics were extracted: literature search terms and time period, geographical coverage, data literature sources, MENA countries with retrieved data, along with the number of included studies, inclusion and exclusion criteria, targeted review population, and reported PA-related outcomes. From each primary study included in a SR, the following characteristics were collected: study design and sample size, sampling method, study setting, years of data collection, population characteristics (type, age, gender, proportion of males and females, and response rate), and all PA-related outcomes (definition, measurement tool/administration, prevalence, and barriers to PA). In case of discordance between the reported data by the SR and data available in the primary study, the latter was retained. A consensus meeting with SC1, KC, and SC2 was held to resolve any disagreements.
In order to assess the methodological quality of the included primary studies and conduct the quantitative analyses, any relevant study characteristics not reported by the SR were extracted from the primary study publication. In addition, any additional data on PA-related outcomes from a MENA country found in an included primary study but not reported by the SR (usually not part of the objectives) was extracted and reported.
Qualitative synthesis. The qualitative synthesis was done at two levels:

Results
A total of seven SRs and 229 primary studies on the epidemiology of PA, physical inactivity, and/or sedentary behaviour in at least one of the 20 selected MENA countries were included in the qualitative synthesis of this overview (Fig. 1). After exclusion of primary studies included in more than one SR (n = 15), primary studies with overlapping data (n = 8), primary studies reporting only sedentary behaviour data or barriers to PA data (n = 65), and those using non-standard definitions for PA or inactivity, excluded from the meta-analyses (n = 59), a total of 82 studies, including 125 PA prevalence measures were considered for the quantitative analyses. characteristics of the included systematic reviews. The seven SRs included in our overview are described in Supplementary, Table S3. The summary of qualitative results, strengths and limitations, research gaps, and recommendations are synthesized in Supplementary, Table S4. PA-related outcomes were found for all the 20 MENA countries. The physical inactivity and sedentary behaviour measures were the primary outcomes in three SRs 54 . The other SRs reported these outcomes only if included in the primary studies on PA. No SR restricted the primary study inclusion to one specific definition of PA. All the identified data was based on the general population. Of note, only Yammine et al. 56 conducted a meta-analysis of prevalence measures of different levels of PA among UAE adolescents. All the included SRs excluded clinical populations and three SR 53,57,58 excluded primary studies based on their methodological quality of sampling and measurement methodology.
Methodological quality assessment of the included systematic reviews and primary studies. No SR reported the list of excluded studies, assessed publication bias, or included the source of funding for the SR itself and of the included primary studies, as per the AMSTAR recommendations 48 (Table 1). Two SRs 56,58 searched for grey literature sources ( Table 1).
The checklist used for the quality assessment of the included studies is presented in Supplementary, Table S5. The quality assessment scores of each reported PA related outcome are reported in Supplementary, Tables S8-S11. The methodological quality scores varied between 11 and 20 out of a maximum score of 21. Most of the reported outcomes had a quality score closer to the upper score limit. Most of the primary studies were cross-sectional clearly describing the population characteristics and study setting, used PA definition consistent with international recommendations, utilized validated measurement tools, used probability-based sampling method, had an equal male and female ratio, and country-level coverage. overview of studies with physical activity data. Forty-seven primary studies with 75 PA prevalence measures on adults and thirty-five primary studies with 50 PA prevalence measures on youth were included in the meta-analysis. The study characteristics and PA prevalence measures among the adult and youth population are summarized in Supplementary, Tables S8 and S9. For both populations, limited data on PA was reported in MENA countries with the exception of the GCC countries (Bahrain, Kuwait, Oman, Qatar, Saudi Arabia and UAE). In most included studies pertaining to GCC countries, the distinction between the national and non-national population was unclear. A wide range of PA prevalence measures were reported for the MENA region. A higher PA prevalence measure among males than females was observed for all included MENA countries except among adults in Lebanon  In most of the MENA countries among both adult and youth populations, more males engage in PA than females except for adults in Lebanon and Jordan (Tables 2, 3, Fig. 2). In MENA as a whole and within GCC countries, greater PA participation was observed among adults than in youth. However, caution must be practiced in interpreting these results given the limited number of included studies and the high heterogeneity between the included prevalence measures.
For adults, most studies used PA definitions consistent with international recommendations of i) at least 150 minutes of moderate-intensity activity per week or ii) at least 600 or more metabolic equivalent of task (MET)-minutes per week of vigorous or moderate activity 6,59,60 . For youth, most studies used a PA definition corresponding to 60 minutes of moderate intensity activity as per the Arab Teens Lifestyle Study (ATLS) questionnaire PA measurement and compilation (Supplementary, Table S12) 61 . www.nature.com/scientificreports www.nature.com/scientificreports/ Only one study 62 used a daily threshold of ≥13,000 steps using the pedometer to quantify sufficient PA participation (Supplementary, Table S12). Most studies pertaining to adults and youth utilized validated tools along with a validated version in Arabic or Urdu language, depending on the primary language of the country.
The results from univariate meta-regression analyses assessing the relationship between study-level covariates and the PA prevalence measures are summarized in Table 4. Factors associated with a higher prevalence of PA among adults were the use of the two validated questionnaires-IPAQ and GPAQ, and data collected after the year 2000. Factors associated with a higher prevalence of PA among youth were male gender and the use of IPAQ. The HBSC, PACE+ questionnaires and country level coverage of the study are associated with a lower PA prevalence among youth ( Table 4). overview of studies with physical inactivity data. To define physical inactivity, we used physical activity levels below the recommended levels of PA. Hence, the validated measurement tools used to assess PA participation were also used to assess physical inactivity (Supplementary, Table S12).
Two studies used a daily threshold of <10,000 steps using the pedometer to quantify insufficient PA participation. Two studies used a daily threshold of a daily heart rate <159 bpm for at least 20 min and a daily heart rate <140 bpm for at least 30 min to quantify physical inactivity.
Among adults and youths, a total of 134 primary studies reported data on physical inactivity (Supplementary, Tables S10 and S11). Studies using a standard definition for physical inactivity and not overlapping with an already included data/study were included in the meta-analysis and are listed in Supplementary, Tables S6 and S7. Overall, higher prevalence measures of physical inactivity among female adults compared to male adults was noted.
A total of 58 primary studies reported factors positively correlated to physical inactivity and barriers to physical activity.
Among the included SRs, only the SR of Ranasinghe 2013 54 reported an OR comparing the prevalence of physical inactivity between males and females in Pakistan. This SR concluded that females were significantly more inactive than males [OR: 2.1 (95% CI = 1.5-3.1), p-value<0.001]. In UAE, Yammine et al. 56 Table S11). The validated tools utilized for sedentary behaviour measurement were the same tools used for PA assessment among adults and youth.

Discussion
Our overview included seven SRs, 229 primary studies and 203,617 participants spanning the 20 MENA countries. Our meta-analysis on PA participation after the year 2000 demonstrates that 49.2% of adults and 74.4% of the youth population were not sufficiently active. Insufficient evidence is available on sedentary behaviour due to the absence of clear measurement guidelines.
Our regional pooled PA prevalence of 50.8% among adults was similar to the recent estimate in the region of MENA and Central Asia (67.2%) 26 but lower than the global estimates of 72.5-77% 25,26 . Minor differences in PA prevalence estimates can be explained by the differences in the number of countries considered for the calculation, the measurement of PA, and data management 25 . Evolving PA recommendations could also be an explanation; previously, sufficient PA was defined as undertaking at least 30 minutes per day of moderate intensity activity for at least five days per week or at least 20 minutes per day of vigorous intensity activity for at least three days per week or an equivalent combination achieving 600 METs-minutes per week 85,86 . This recommendation is likely more difficult to be achieved when compared to the updated recommendation of 150 minutes of moderate intensity activity or 75 minutes of vigorous intensity activity per week, or an equivalent combination regardless of the weekly frequency 6,7 , since this has more flexibility.
Our regional pooled PA prevalence of 25.6% among youth is consistent with the global estimate of about 20% of youth who are sufficiently active as per the international recommendation for PA 9,25 . The lack of awareness pertaining to the recommendation of 60 minutes of moderate-to vigorous-intensity PA daily for children and youth 6,7,87 and its benefits can explain the very low PA participation in the region 57,58 and globally. Our findings in MENA confirm the lower PA participation of youth in comparison to adults which is observed globally 9,25 . This could partially be due to the higher recommended levels of PA for youth and in part due to the tools utilized to measure PA. Because PA habits developed during youth may persist in adulthood, young people should be encouraged to participate in a variety of physical activity that supports their natural development 6 .  Table 2. Meta-analysis of physical activity prevalence in MENA countries among the adult general population. Notes: Mean effect size reported as weighted average prevalence measure with their corresponding 95% confidence interval (CI); the p-value of the Cochran's Q statistic is used to inform about the statistical significance of the heterogeneity in PA prevalence estimates within the group; the p-value of the Cochran's Q between-subgroups statistic is used to inform about the statistical significance of the heterogeneity in PA prevalence estimates between the results of subgroups if applicable; the I 2 is used to assess the magnitude of between-study variation that is due to differences in PA prevalence estimates across studies rather than chance; the prediction interval is used to estimate the 95% interval in which the true PA prevalence in a new PA prevalence study will lie. No meta-analyses according to the time of data collection were then performed. p-value <0.05 was considered statistically significant. Meta-analyses according to the type of population (nationals and non-national populations) were only relevant in the GCC countries as per the type of setting of the GCC population. Years of data collection were considered as prior to the publication year in studies where this information was missing. The prevalence measures were then classified accordantly for the meta-analyses. <2000: this analysis includes all studies where the data was collected before 2000.   www.nature.com/scientificreports www.nature.com/scientificreports/ Implementation of locally informed, evidence-based interventions promoting PA participation at the standardized recommended levels and addressing barriers for PA would be a step in the right direction.
Our findings confirm the higher PA participation among males as compared to females which is observed globally 9,22,88,89 and regionally 52,53,56-58 among both adults and youth. Hence, the female youth should be a target population for future interventions to increase PA participation in the region. Understanding the barriers and determinants of PA in the MENA population is essential for developing culturally-and age-appropriate PA interventions ( Table 5). The sex differences in adults may be attributable to the sociocultural role of men and women in the region 54,58 , however, barriers related to the lower participation of female youth in PA requires further investigation (Table 5). Physical activity interventions should be inspired from countries with the highest PA prevalence measures among girls, for example in China (75%) 88 , where walking and biking are a common form of commuting. Our analyses provide some indication of an increased PA participation among adults after 2000 in all MENA countries with available data in both periods. Although definitions used for sufficient PA participation were similar between the studies conducted before and after 2000, caution must be practiced while interpreting these results given the limited data collected before 2000. In addition, none of the studies conducted before 2000 assessed PA participation using a standard PA questionnaire (GPAQ, IPAQ) as these questionnaires were developed in 2002 90 and 2003 91 respectively.
Unlike other MENA countries, a substantial proportion of the population in the GCC countries is non-national 37 . Physical activity data differentiating nationals and non-nationals is scarce. The increase in PA in the GCC after 2000 may be explained by urbanization and socioeconomic improvement, see Table 5. Additional studies considering the demographic specificities in the GCC are needed to provide evidence on PA behaviour to develop, implement, and monitor public health programs.
We found a high proportion of heterogeneity between PA prevalence measures attributable to the type of questionnaire used. Prevalence estimates using the IPAQ were higher than those using the GPAQ. This is consistent with previous findings demonstrating PA over-reporting using the IPAQ compared to other similar questionnaires such as the GPAQ 27 , the Behavioural Risk Factor Surveillance System (BRFSS) 92 , or the accelerometer 93,94 . This results in an overestimation of the PA prevalence measures and consequently an underestimation of the physical inactivity prevalence measures 95 . Moreover, the GPAQ and IPAQ measure the total PA for a typical week making comparison difficult with the BRFSS (non-occupational PA for the past month) 96 and the Active Australia Survey (leisure-time PA in the week preceding interview) 97 used in Western countries.
The GPAQ is a valid measurement tool to assess moderate-to-vigorous PA endorsed by the World Health Organization 90 . However, this tool demonstrated a weak agreement with accelerometer use in young Saudi men 98 . Adapting the current GPAQ to build a regional PA questionnaire has been recommended 98 . Both measurement tool-based factors (question phrasing, time period coverage, and single or multipart questions) 99 and individual-based factors (demographic and socioeconomic characteristics) 98 should be considered in its development.
The international PA recommendations for youth require a mix of moderate and vigorous intensity activity (vigorous-intensity activity, including those that strengthen muscle and bone, at least three days per week) 6,7,87 . Physical activity participation among youth assessed using the ATLS questionnaire was based on the conversion of all moderate and vigorous activities during the week to a summary measure in MET-minutes with a cut-off of 1,680 METs-minutes per week corresponding to 60 minutes of moderate intensity activity every day of the week 61 . This methodology of quantifying PA participation is inclusive of participants engaging in solely moderate activities while the standard definition includes only those participants who engage in a combination of moderate activities most days of the week and vigorous activities three times per week 6,7,87 . The ATLS questionnaire could then overestimate the proportion of the population which is physically active. Nevertheless, the ATLS questionnaire was the only PA questionnaire validated among youth in MENA countries.
Objectively measured PA using accelerometer, pedometer or continuous heart rate monitor can avoid recall bias and is practical for use particularly among youth when compared to self-reported data 100 . However, the minimum number of steps measured by the pedometer required for health benefits is still debatable [101][102][103][104] and no significance of the heterogeneity in PA prevalence estimates within the group; the p-value of the Cochran's Q between-subgroups statistic is used to inform about the statistical significance of the heterogeneity in PA prevalence estimates between the results of subgroups if applicable; the I 2 is used to assess the magnitude of between-study variation that is due to differences in PA prevalence estimates across studies rather than chance; the prediction interval is used to estimate the 95% interval in which the true PA prevalence in a new PA prevalence study will lie. p-value <0.05 was considered statistically significant. <2000: this analysis includes all studies where the data was collected before 2000. > 2000: this analysis includes all studies where the data was collected before 2000. Meta-analyses according to the type of population (nationals and non-national populations) were only relevant in the GCC countries as per the type of setting of the GCC population. Years of data collection were considered as prior to the publication year in studies where this information was missing. The prevalence measures were then classified accordantly for the meta-analyses. All data among youth was collected after 2000. Symbols: * Include converted data from physical inactivity to physical activity not overlapping with other included data. **include data from an included primary study not reported by the SR. *** Data on moderate and vigorous physical activity levels were merged in the studies of Wasfi, 2008 36 andMehairi, 2013 34 , respectively, to ovoid partial overlapping of denominators data from the same study. a Stratified prevalence measures were used to conduct the meta-analyses. b Weighted average prevalence measures were obtained using random-effect model. c Excluded the studies with a mixed male and female population. (2020) 10:9363 | https://doi.org/10.1038/s41598-020-66163-x www.nature.com/scientificreports www.nature.com/scientificreports/ health-based criteria for steps per day have been developed for adults and youth 88,100,102,104,105 . Despite attempts to establish the steps per day threshold equivalent to the daily recommended PA level for youth 106   www.nature.com/scientificreports www.nature.com/scientificreports/ guidelines use the steps count to define sufficient PA 6,7,49,86 . Additionally, even if it was used as a gold standard in the ATLS validation study 109 , the pedometer is not designed to capture pattern, intensity, or type of PA and most are sensitive to ambulatory activities 105 .
The most basic dimensions of PA assessment are frequency, intensity, time (duration), and type of activities 100 . Currently, no single tool can adequately assess all these dimensions 101 . A combination of self-reported measures (to capture important domain-and behaviour-specific sedentary time information) and device-based measures (to measure both total sedentary time and patterns of sedentary time accumulation) is recommended 99,110 .
Sedentary behaviour is a different concept than physical inactivity and is expressed as a total number of minutes or hours spent in sedentary activities on a typical day 60 Table 4. Univariate meta-regression models for physical activity prevalence in MENA adult and youth general population. Notes: Significant results are highlighted in bold. Association between the meta-regression variable and the prevalence of PA is reported as an odds ratio (OR) with its corresponding 95% confidence interval. Mean effect size reported as weighted average prevalence measure with their corresponding 95% confidence interval (CI). Weighted average prevalence measures were obtained using random-effect model. Proportion of total between-study variance explained by the model (R 2 ) is used to assess the magnitude of the covariates relationship with effect size 30 . All data among youth was collected after 2000. No meta-analyses according to the time of data collection were then performed. The category "Country level" includes all prevalence measures based on a national sample of individuals representative of the entire country where the study was conducted (e.g. all cities of the country). To be included in this category, a study should clearly mention that the study was conducted at the national level. Any study based on a sample from a restricted part of the country (one or few cities, specific university, specific school, etc.) is considered in the "Local level" category. † p-value <0.05 was considered statistically significant. www.nature.com/scientificreports www.nature.com/scientificreports/ considered in the sedentary behaviour questionnaires prevent comparisons between countries in the region and globally. To date there is no well-accepted threshold to categorize sedentary behaviour data 59 .
The accelerometer is acknowledged as a valid and reliable instrument for objective measurement of sedentary behaviour in adults and youth [111][112][113][114][115] . However, it is a costly option for researchers. Moreover, it cannot provide information on the type and setting of sedentary behaviour and there is no consensus on the number of steps to identify individuals with sedentary behavior 110,116 . Sedentary behaviour has been recognized as a public health issue only in the past decade 9 and therefore, only minimal standardized instruments exist for its assessment with no clear interpretation guidelines.
To our knowledge, this is the most comprehensive systematic overview and meta-analysis covering several dimensions of PA behaviour in the MENA region. Our overview is the first to bring together the available evidence on sedentary behaviour and PA. Moreover, most of the included primary studies have assessed PA prevalence using validated measurement tools in random national-level samples supporting the quality of the reported estimates. The country-, time-, age-, and sex-specific pooled PA prevalence measures are a major addition to the evidence on PA. This compilation will serve as a benchmark for epidemiologists and public health interventionists (Table 6). This overview identifies barriers (Table 5) and research gaps ( Table 7) that could help direct future funding and investigation.
Limitations include the restriction of the search strategy to PubMed/MEDLINE. However, we searched additional sources indexing grey and non-grey literature relevant to the region. Moreover, the included SRs have searched in total 21 literature sources for primary studies including grey literature sources, which are potentially important data sources for MENA countries 27,31 . This minimizes the risk of publication bias in our overview. We also extracted the data from the primary studies to complete the missing information and provide a complete overview of the available data.
The data that we used on PA and sedentary behaviour was collected before 2018; new PA data from MENA countries may have been published since. The limited data on PA behaviour in the MENA countries could be explained by the targeted population and outcomes of the included SRs. To optimize the quantitative synthesis, we converted physical inactivity data, using a standard definition, to PA measures. Despite attempts to minimize the heterogeneity using several subgroup analyses, a high heterogeneity between the PA prevalence measures persisted. Although most of the primary studies used validated PA-measurement tools, differences in PA domains captured by these measurement tools contributes to the heterogeneity between estimates. Another limitation of our metanalysis could be the minor differences in the definitions of the 'active' category among youth (e.g. number of days per week or steps). However, since only a limited number of studies had these differences, we feel this may have a minimal impact on the results. Differences according to the area of residence (rural versus urban) and the study coverage (country versus local coverage) might have been additional sources of heterogeneity in the reported prevalence measures.
• The employment of domestic helpers 53 .
• Sociodemographic factors (eg: Age; Sex; Education level) 62  www.nature.com/scientificreports www.nature.com/scientificreports/ In comparison to the global average estimates, our regional PA prevalence estimates for MENA were lower among both adults and youth. The lower PA participation by females and youth observed globally was also found in several MENA countries. More recent data suggests an increased PA participation in the region as a whole. Some of the variability in the PA measures might be attributable to the type of questionnaires used for data collection. Trends and patterns need to be confirmed in future investigations. Whether sedentary behaviour is a public health issue in the MENA region remains unclear, due to the absence of guidelines to quantify and interpret it. Harmonizing definitions of sedentary behaviour would be useful to generate evidence for the region. Promoting PA and limiting sedentary behaviour should be health priorities for both adults and the youth to reduce the overall burden of obesity and other non-communicable diseases. Additionally, promoting PA surveillance and identifying and investigating sedentary behaviour trends utilizing validated and reliable measurement tools consistent at national and international levels is essential to allow meaningful comparisons and to implement effective evidence-based interventions.

Data availability
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request. • Public health efforts to increase PA and decrease sedentary time require standardized PA surveillance among all countries in the region. These measures are necessary to understand locally informed interventional strategies and to identify target populations affected by physical inactivity. This requires building and strengthening country capacity with a systems approach to scale up efforts aimed at increasing PA.
• Health-care providers can adopt a comprehensive curriculum that potentially closes the gap in medical schools, residency programs, graduate education, and nursing curricula on topics related to PA, and exercise prescription and lifestyle health 57 .
• Current data suggests that MENA youth population is insufficiently active. Physical environment interventions like parks, playgrounds, sports venues, proximity to destinations, and neighbourhood aesthetics can be beneficial. Additionally, psychosocial factors such as self-efficacy and social support from peers should also be considered while planning interventions.
• Identifying and considering PA-relevant domains (e.g. transportation, leisure-time PA) and barriers specific to the MENA populations are essential for adapting national actions and policies to tackle obesity and other non-communicable diseases.
• Considering the increasing urbanization in the region 115 , urban design favourable for PA has been suggested as a potential strategy to mitigate the impacts of urbanization on PA in low-and middleincome MENA countries 25 .
• Clear guidelines for sedentary behaviour measurement and interpretation in order to characterize patterns of sedentary behaviour in the MENA countries are required. Table 6. Public health implications and recommendations.
There is a need to: • Clarify standardized methods of measuring PA, inactivity and sedentary behaviour for future studies.
• Generate data on PA-related outcomes in MENA countries other than the GCC countries.
• Explore differences in PA participation between nationals and nonnationals in the GCC countries.
• Further understand the personal, social and environmental barriers to PA, particularly in relation to the different domains of PA (leisure time, occupational, transports, and households PA) specific to the region 53,56 to facilitate effective locally informed interventions.
• Follow PA trends over time 53 . Table 7. Evidence gap and future research.