Associations of Physical Activity and Sitting Time With the Metabolic Syndrome Among Omani Adults

Authors


(rmmabry@gmail.com)

Abstract

Most findings on associations of physical activity and sedentary behavior with the metabolic syndrome are from developed countries; thus, we examined these relationships in adults from Sur, Oman. The Sur Healthy Lifestyle Survey (n = 1,335) used the World Health Organization (WHO) Stepwise methodology to assess chronic disease risk factors. Odds ratios for the metabolic syndrome were estimated using logistic regression models for domains of physical activity (work, transport, and leisure) and sitting time, and adjusted for confounding variables. Compared to their counterparts doing the least physical activity, lower odds of the metabolic syndrome were observed among those with higher work activity (0.60; 95% confidence interval (CI): 0.45, 0.80) and transport activity (0.69; 95% CI: 0.47, 1.00), but not leisure activity (0.91; 95% CI: 0.64, 1.32). Odds of the metabolic syndrome were higher in those who sat for ≥6 h daily compared to <3 h daily (odds ratio = 1.60, 95% CI: 1.04, 2.44), but not after further adjustment for physical activity. This is the first evidence from the Arabian Gulf on associations of physical activity and sitting time with the metabolic syndrome and provides empirical evidence to inform national physical activity guidelines, policies and programs.

Introduction

The significant association of physical activity (primarily in leisure time) with the metabolic syndrome, with an increased likelihood in those who do not meet physical activity recommendations compared to their more active counterparts, has been reported widely (1,2,3,4). These associations tend to be stronger in men than in women (2,3). Protective effects of physical activity on the metabolic syndrome have been demonstrated in prospective studies (4,5). Sedentary behavior (too much sitting as distinct from too little exercise) is a more-recently identified health risk: it has been found to be associated with the metabolic syndrome and with its components in both cross-sectional (2,3,6) and prospective studies (7,8).

The majority of findings on associations of physical activity and sedentary behavior with the metabolic syndrome have come from developed countries, including Australia, the USA, and some countries in Western Europe (1,2,3,4). No such evidence is available from the countries of the Gulf Cooperation Council (GCC; Bahrain, Kuwait, Oman, Qatar, Saudi Arabia, and United Arab Emirates). The rapid socioeconomic development in the GCC in recent years has been accompanied by an epidemiologic transition including rising prevalence of obesity and diabetes (9,10). The prevalence of the metabolic syndrome (a combination of the metabolic precursors of chronic diseases) is 10–15% higher compared to developed countries, and is generally higher in women than men (9). The profile of relevant behavioral risk factors is different compared to the western populations from which most of the evidence has been derived. For example, from studies based on self-reported physical activity, the prevalence of adults being physically active in the GCC is low, ranging from 39 to 42% for men and 26 to 28% for women, compared to western populations where it ranges from 50 and 60% for men, and 47 and 54% for women in the USA and Australia, respectively (11,12,13). This suggests that the epidemiology of such behaviorally driven chronic diseases needs to be examined specifically within GCC populations in order to provide evidence that can inform the relevant preventive health programs and policy initiatives. We, therefore, examined the associations of physical activity and sitting time with the metabolic syndrome among Omani adults and whether the associations differed by sex.

Methods and Procedures

Overview of the Sur Healthy Lifestyle Survey

The Sur Healthy Lifestyle Survey is a cross-sectional survey conducted in the city of Sur (on the north-east coast of Oman, 150 km from Muscat, the nation's capital) in 2006, as a baseline study for the Sur Healthy City Project. This survey followed the World Health Organization (WHO) Stepwise methodology, a standard method of data collection on key risk factors for chronic diseases (14). The study sample comprised Omani men and women aged 20 years or older residing in Sur city.

A proportional-to-size sampling design was used to ensure that the selected sample was representative of the city as a whole. A random sample of 1,700 houses was selected from 191 clusters within 9 census enumeration areas. Within each selected house, an individual was randomly selected from all eligible household members (≥20 years and nonpregnant). The household interview and all laboratory tests were completed by 1,373 individuals, giving a response rate of 80.8%. Ethical approval was granted by the Ministry of Health Research Committee for the Sur Healthy City Project survey. The present analyses used data from all participants with complete data for all components of the metabolic syndrome, physical activity, sitting time as well as the identified confounders (n = 1,335).

Survey instrument

The Arabic version of the WHO Stepwise instrument was used. This instrument collects cardiovascular clinical and behavioral risk factors, demographic information and family history of chronic diseases (described below). The Physical Activity Module uses the Global Physical Activity Questionnaire (GPAQ), an instrument with established validity and reliability (15). GPAQ is a 16-item instrument developed by the WHO to measure physical activity (intensity, duration, and frequency) performed in three domains—work (paid and unpaid, including housework), transport (walking and cycling) and leisure—as well as total sitting time, across diverse populations (15,16). However, validity and reliability testing did not include populations from the Arab world.

Data collection procedures

Data were collected in two steps: face-to-face household and individual interviews followed by a visit by the selected individual to the closest health centre (after an overnight fast) for the physical and clinical measures. Demographic (sex, age, marital status, education, and work status), behavioral risk factors (tobacco use, fruit, and vegetable intake in addition to physical activity and sitting time) and family history of chronic disease indicators (heart attack, stroke, diabetes, cancer, hypertension, and high cholesterol) were obtained from study participants during the individual household interviews. In addition, participants were asked if they had ever been diagnosed with having diabetes and hypertension and if so, if they were taking medications or other treatment.

During the health center visit, two blood specimens were collected from all participants for the assessment of blood glucose, cholesterol, and triglycerides and were analyzed at a central laboratory. Samples for estimation of glucose level were collected in sodium fluoride oxalate anticoagulant container (red) and samples for estimation of lipid profile were collected in a plain container (red). After plasma separation by centrifuge, glucose and lipids assessment was conducted in a Hitachi 902 automated clinical chemistry analyzer (Roche, Basel, Switzerland) by enzymatic colorimetric method. Two blood pressure readings 3 min apart were taken in a seated position using a portable sphygmomanometer; with the mean of the two readings used. Anthropometric measurements (height, weight, and waist circumference) were taken following the WHO protocol (14). Data entry was carried out using SPSS (v.9) after questionnaires were reviewed for quality checks.

Dependent variable—the metabolic syndrome

Using one of the recent internationally-recognized definitions, the National Heart, Lung, and Blood Institute/American Heart Association criteria (17), participants were identified as having the metabolic syndrome if they had at least three of the following (or were taking appropriate medication): central obesity using ethnic-specific thresholds for the Arab population (≥94 cm for men and ≥80 cm for women), raised triglycerides (≥1.7 mmol/l), reduced high-density lipoprotein-cholesterol (<1.0 mmol/l for men or <1.3 mmol/l for women), raised blood pressure (systolic blood pressure ≥130 mm Hg, diastolic blood pressure ≥85 mm) or raised plasma glucose (fasting plasma glucose ≥5.6 mmol/l).

Independent variables—physical activity and sitting time

Time spent in each domain of physical activity was multiplied by metabolic equivalents (MET) to yield MET-minutes (MET-min). MET is the ratio of specific physical activity metabolic rates compared to the resting metabolic rate, where one MET is equivalent to the energy cost of sitting quietly (1 kcal/kg/h). Based on the GPAQ analysis guidelines, for all domains moderate-intensity activities were assigned a value of 4 METs; vigorous-intensity activities were assigned a value of 8 METs. The total MET-min were computed as the sum of all MET-min/week from moderate-to-vigorous-intensity physical activities performed for each domain (18). Associations of physical activity with the metabolic syndrome were not linear; therefore, each domain-specific physical activity variable was examined as categories. Evenly spaced categories, representing multiples of 30 min of moderate activity at 3.5 METs intensity (i.e., 105 MET-min) were initially examined. Based on the shape of the associations seen in the bivariate analysis (not presented), physical activity variables then were collapsed into the following categories for analysis: work (≤315 MET-min/day, 315+ MET-min/day), transport (0, >0–105 MET-min/day, 105+ MET-min/day) and leisure (0, >0–105 MET-min/day, 105+ MET-min/day). Overall reported sitting time per day was categorized as <3 h/day, 3– <6 h/day, and 6+ h/day.

Data analysis

Data analyses were carried out using STATA (v.11). We described the sample overall and separately by sex (without weighting). Data for bivariate and adjusted analyses were weighted to the age and sex distribution of the Sur city population using weights calculated from the National Census 2003 (19). Statistical significance was set at P < 0.05 and P < 0.1 for interactions (both two-tailed). Associations of the three domains of physical activity and sitting time with the metabolic syndrome were estimated using logistic regression: firstly without adjustment; secondly, adjusted for potential confounding variables (sex, age, marital status, education, work status, tobacco use, fruit and vegetable intake, and family history of heart attack, stroke, diabetes, cancer, hypertension, and/or high cholesterol). The relevant confounding variables were identified using backward elimination, in pooled models and separately by sex. Initially, demographic, behavioral risk factors, and family history of chronic disease indicators were considered as potential confounding variables. Variables that were associated with the metabolic syndrome at P < 0.2 in either pooled or sex-stratified models were retained as potential confounders. To ensure the association of sitting time was also independent of physical activity an additional model for sitting time was run that further adjusted for any physical activity domains that showed an association with the metabolic syndrome. Results from these models were reported as odds ratios with 95% confidence intervals (CIs); marginal probabilities were also used to report on adjusted prevalences of the metabolic syndrome, separately for men and women. To examine whether associations of physical activity and sitting time with the metabolic syndrome differed by sex, interactions were tested without weighting to avoid inflated standard errors.

Results

The sample from Sur city, Oman was relatively young with a mean age of 36.4 years with about half having at least secondary-level education (Table 1). One out of five men were current smokers (21.0%), while very few women smoked (0.3%). The median number of servings of fruits and vegetables eaten per day was very low (1.86 servings/day, data not shown) compared to the five or more servings recommended by the Ministry of Health, Oman (20).

Table 1.  Selected sample characteristics of men and women from the Sur Healthy City Survey
inline image

The percentage of the sample with the metabolic syndrome was 27.3% and was higher in women (29.8%) than in men (24.2%). The percentage identified as overweight/obese was 65.0%: twice as many women were obese compared to men (41.8% for women and 21.4% for men) with BMI as the indicator. While the mean waist circumference was similar for men and women (88.7 cm, SD 13.2, and 89.2 cm, SD 15.4, respectively), the percentage classified as obese based on waist circumference was also twice as high for women (72.7%) than for men (35.0%).

Work physical activity was the most commonly reported activity domain, with nearly two-thirds (61.3%) of the sample reporting doing at least 316 MET-min/day. However, nearly one-third (30.5%) reported not doing any transport physical activity and more than half (55.4%) reported not doing any leisure physical activity. Women reported more work physical activity than did men, whereas men reported more transport and leisure physical activity than did women. Almost twice as many men (17.8%) as women (9.4%) sat for six or more h/day.

Table 2 shows associations of domain-specific physical activity and sitting time with the metabolic syndrome. There were no significant sex interactions in associations of physical activity and sitting time with the metabolic syndrome (all P > 0.1). Thus, only the pooled models are reported. Bivariate analyses showed that for each domain of physical activity, the percentage of participants with the metabolic syndrome tended to be lower at higher levels of physical activity, with both work and leisure physical activity showing significant associations. On the other hand, higher levels of sitting time were associated with higher levels of the metabolic syndrome, but the association was not statistically significant (P = 0.134).

Table 2.  Associations of physical activity and sitting time with the metabolic syndrome among adults in Sur city, Oman (N = 1,335)
inline image

In adjusted analyses, compared to their counterparts doing the least physical activity, lower odds of the metabolic syndrome were seen for those with higher work physical activity (significant) and transport physical activity (borderline). Leisure physical activity did not show an association. Sitting time was significantly associated with the metabolic syndrome when adjusted only for the identified confounders; those who sat for six or more h daily had a one and a half fold higher odds of the metabolic syndrome than did those who sat for <3 h (odds ratio = 1.55, 95% CI 1.01, 2.38, P < 0.05; data not shown). However, after further adjustment for work and transport physical activity, the association of sitting time with the metabolic syndrome was attenuated and no longer statistically significant.

Figure 1 shows the estimated prevalence of the metabolic syndrome in men and women in Sur city by physical activity and sitting time after controlling for confounders. The prevalence rates were higher for women than men across all levels of physical activity and sitting time, however, the amount of difference varied across domains. The highest prevalence rates were seen in women who sat >6 h daily (33.9%) and women who did low work physical activity (32.8%). The lower prevalence rates were seen in men with high levels of work physical activity (13.7%).

Figure 1.

Estimated prevalence of men and women with the metabolic syndrome among residents of Sur city, Oman, by domains of physical activity and overall sitting time (adjusted for confounders). Work: low (<315 Met-min/day), high (316+ Met-min/day); transport: none, low (>0–105 Met- min/ day), high (>106 Met-min/day); leisure: none, low (>0–105 Met-min/day), high (>106 Met-min/day); sitting time: low (<3 h/day), medium (3–6 h/day), high (>6 h/day). Rates weighted to the Sur city population (2003) and adjusted for age, education, marital status, weekly fruit and vegetable intake, family history of chronic diseases (stroke, diabetes, cancer, hypertension, and high cholesterol). Error bars denote confidence intervals.

Discussion

This is the first evidence from the Arabian Gulf Region on associations of physical activity and sitting time with the metabolic syndrome. One out of four people in the sample population had the metabolic syndrome, with higher rates in women than in men (29.8 and 26.4%, respectively). The trend of higher rates in women is also seen in the Oman population as a whole (women: 40.0%, men: 18.4%) as well as in other countries of the GCC (Qatar, women: 37.7%; men: 29.6% and United Arab Emirates, women: 45.9%; men: 32.9%) (9). However, the opposite is true for rates reported for Western countries like Australia (women: 15.7% and men: 26.4%) and the USA (women: 21.4% and men: 26.9%) where the prevalence of the metabolic syndrome is higher in men compared to women (21,22). The high birth rate (mean number of births for the A'Sharqiyah Region is 5.4) may partially explain the high rates of central obesity and thus, higher rates of the metabolic syndrome among women (9,23).

Lower odds of the metabolic syndrome were seen for those with higher work and transport physical activity compared to their least active counterparts, but not for leisure physical activity. The beneficial association of physical activity with the metabolic syndrome found in this study is similar to findings from other population groups in both developed and developing countries (1,2,3,24,25). However, studies from developed countries typically have examined the association of leisure physical activity (sometimes including transport) and the metabolic syndrome (2,3,4). Our study is one of the few to examine the association of different domains of physical activity with the metabolic syndrome; this is an important consideration, particularly for countries like Oman and Saudi Arabia where work and transport physical activity contribute more to total physical activity than does leisure physical activity (26).

The lack of a significant beneficial association between leisure physical activity and metabolic syndrome is a finding not usually seen in studies in developed countries. There are many unique aspects in the Sur city population that may contribute to this, including the low rates of leisure physical activity. However, given previous prospective-study evidence showing protective effects of physical activity on the metabolic syndrome (4,5), greater efforts to increase physical activity in the study population would likely be beneficial. Further research is required to better understand domain-specific physical activity in Oman and its relationship to the metabolic syndrome and other health outcomes.

Significantly higher odds of the metabolic syndrome were seen in those who sat for six or more h daily compared to those who sat for <3 h/day, but not after adjustment for work and transport physical activity. As with studies from developed countries (1,2,3) as well as in some developing countries such as Iran (24), we saw an adverse association of sitting time with the metabolic syndrome; however, unlike these other studies, our associations were not independent of physical activity as associations were not present after adjustment for work and transport physical activity.

Since the data were weighted to the Sur population and the study had a high response rate (80%), the findings are likely to reflect the situation of adult residents of Sur city at the time of the study. This generalizability is limited, however, by the fact that information regarding nonrespondents was not available at the time of analysis. The findings cannot be generalized to the overall Oman population since this study was limited to a specific region of the country. Also, the prevalence of the metabolic syndrome in women is much higher in Oman as a whole than it is in Sur (27). The findings also cannot be generalized to the GCC as a whole since the prevalence of the metabolic syndrome in Sur is lower than in other countries of the GCC (9). Due to the cross-sectional nature of this study, the associations of physical activity and sitting time with the metabolic syndrome do not necessarily indicate a causal relationship. Given the limitations of measuring physical activity by self-report (28) and the fact that the validity and reliability of the GPAQ instrument has not been tested for populations in the Arab world (15), it is possible that some individuals were misclassified.

The adjusted prevalence of the metabolic syndrome according to physical activity domains and sitting time levels helps identify high risk groups, particularly women who do less than 315 MET-min of work physical activity a day, no transport physical activity or who sit for >6 h daily. These findings highlight the need for more policy level attention to physical activity and sedentary behavior. Although international guidelines are readily available (29,30), public health policies and programs need to focus on improving physical activity within relevant sociocultural contexts. For example, in Oman and similar cultures, policies and programs should take into consideration the low value placed on physical activity, social expectations of gender segregation (i.e., gender-specific exercise facilities), restrictions of women's freedom of movement outside the home and the common practice of employing expatriate domestic workers (31,32,33,34). Further research to understand the characteristics of those who are least active or most sedentary is needed to assist in targeting appropriate interventions. The alarmingly high rates of obesity in this population compared to Western countries, particularly for women (10,32,35,36), may pose further barriers to physical activity that are additional to the likely roles of sociocultural norms that may limit women's options. Culturally specific policy and programs are therefore warranted.

ACKNOWLEDGMENTS

We thank the Director General of Health Service, A'Sharqiyah Region, Ministry of Health, Oman and those involved with the Sur Healthy City Project for their willingness in sharing their data with us. E.A.H.W, E.G.E, M.M.R. and N.O. are supported by Program and Fellowship Grants from the National Health and Medical Research Council of Australia and by Research Infrastructure funding from Queensland Health.

DISCLOSURE

The authors declared no conflict of interest.

Ancillary