Journal Logo

BASIC SCIENCES: Epidemiology

Physical Activity, Obesity Status, and Glycemic Control

The ATTICA Study


Author Information
Medicine & Science in Sports & Exercise: April 2007 - Volume 39 - Issue 4 - p 606-611
doi: 10.1249/mss.0b013e31803084eb
  • Free


The incidence of metabolic diseases has escalated dramatically worldwide in recent years, highlighted by the fact that both diabetes mellitus and obesity have reached levels of uncontrollable epidemics. According to the World Health Organization (WHO) estimations, the total number of people with diabetes in year 2000 was 171 million, and it is projected to rise to 366 million in 2030 (26). In parallel, the prevalence of obesity in the United States in 1999-2000 was 30.5%, compared with 22.9% in 1988-1994 (7).

The basic metabolic abnormality of both diabetes and obesity is resistance of peripheral tissues to the action of insulin. In populations with genetic susceptibility to diabetes, insulin resistance has been characterized as a precursor of diabetes mellitus (12). Obesity has been considered a major risk factor for the development of non-insulin-dependent diabetes mellitus. In a cohort of nearly 7000 men with no history of diabetes, the risk after a mean follow-up of 12 yr for those with BMI levels of 25.0-27.9 kg·m−2 was more than double that for those with BMI < 25.0 kg·m−2, and it increased further with increasing BMI (23). In the same study, weight gain during the follow-up period was associated with a substantial increase in the risk of diabetes, whereas weight loss had the opposite effect. Even though obesity-related insulin resistance may never develop to the typical form of diabetes, the condition itself has been recognized as an independent morbidity risk factor, especially for cardiovascular diseases (18,20).

Epidemiologic evidence suggests that regular physical activity is a key factor in the prevention or treatment of metabolic diseases. Prospective cohort studies have revealed that relatively high physical activity levels are associated with substantial reductions in the risk of developing non-insulin-dependent diabetes mellitus (9), latent autoimmune diabetes (2), and coronary heart disease (21). In addition, extended clinical trials that incorporated a feasible goal of increased physical activity in individuals with impaired glucose tolerance have reported decreases in the risk of developing diabetes by 46-58% after follow-up evaluations of 3-6 yr (6,11,15,22).

To the best of our knowledge, few data on the relationship between estimates of insulin resistance and physical activity levels have been presented. One study of 5159 middle-aged men has shown an inverse relationship between physical activity and fasting insulin levels, with those individuals performing vigorous activities having the lowest insulin levels, despite no differences in blood glucose levels among category groups of physical activity (24). The purpose of the present study was to evaluate the relationship between physical activity status, obesity or overweight, and indices of glycemic control and insulin resistance.


The ATTICA study is a cross-sectional health and nutrition survey carried out in the province of Attica (including 78% urban and 22% rural areas), where Athens, Greece is a major metropolis. The sampling was random, multistage by city, and stratified by age and gender group according to the gender-age distribution of the province of Attica (2001 census). The study's design anticipates enrolling only one participant per household. The main goals of the ATTICA study were to record the distribution of several blood lipids and inflammatory, oxidation, coagulation, thrombotic, and clinical factors, and to explore the associations between these factors with several sociodemographic, lifestyle, and psychological characteristics of the participants.


From May 2001 to August 2002, 4056 inhabitants from the Attica area were randomly asked (via male or telephone) to participate in the study. Of these, 3042 agreed to participate. The participation rate was 75% and did not differ between urban and rural regions. Most of the people who did not participate in the study cited a lack of time. The only information we have about nonresponders concerns age, sex, and region. No differences were observed between responders versus nonresponders regarding these parameters. Each participant gave informed written consent, and the protocol was approved by the review board of the Medical School of Athens. Fifteen hundred fourteen of the participants were men (50%, 20-87 yr old), and 1528 were women (50%, 20-89 yr old). There were 48% men and 52% women between 40 and 49 yr old. The number of participants was determined by power analysis and chosen to evaluate two-sided differences between the normally distributed investigated parameters and physical activity groups greater than 10%, achieving statistical power greater than 0.80 at a P value < 0.05. People living in institutions were excluded from the sampling. The participants had no clinical evidence of cardiovascular or any other atherosclerotic disease, or of chronic viral infections, dental problems, or any type of surgery in the past week. They also exhibited no signs of cold, flu, or any acute respiratory infection.

Physical Activity Ascertainment

We used a translated version of the validated International Physical Activity Questionnaire (IPAQ), suitable for assessing population levels of self-reported physical activity (4). The short version of IPAQ (nine items) that we used provided information on weekly time spent walking at vigorous- and moderate intensity and in sedentary activity. Participants were instructed to refer to all domains of physical activity. Both continuous and categorical indicators were assessed from IPAQ. The continuous indicator was expressed as MET-minutes per week, whereas the categorical analysis grouped the subjects in the following three levels that were developed according to a key concept in current public health guidelines for physical activity (17). The continuous indicator was estimated as a sum of weekly MET-minutes of walking and vigorous- and moderate-intensity exercise. MET-minutes from each category of physical activity were derived from the following formulas:

The factors 3.3, 8.0, and 4.0 were the MET values for walking, vigorous-intensity activity, and moderate-intensity activity, respectively (1). Participants also were classified as inactive, minimally active, and HEPA active (health-enhancing physical activity; a high active category) according to the following criteria. An individual was classified as inactive, which is the lowest physical activity level, when no criteria were met to classify him or her in any of the other two categories. Participants were classified as minimally active, which is the classification for sufficiently active, when any of the following three criteria were met: a) three or more days of vigorous activity of at least 20 min·d−1, b) five or more days of moderate-intensity activity or walking of at least 30 min·d−1, or c) five or more days of any combination of walking, moderate-intensity activity, or vigorous-intensity activity achieving at least 600 MET·min·wk−1. Volunteers were classified as HEPA active when any of the following criteria were met: a) vigorous-intensity activity on at least 3 d, achieving a minimum of at least 1500 MET·min·wk−1, or b) seven or more days of any combination of walking, moderate-intensity activity, or vigorous-intensity activity achieving a minimum of at least 3000 MET·min·wk−1. Participants were instructed to report only episodes of activities of at least 10 min, because this is the minimum duration required to achieve health benefit (17). Values below 10 were recoded to 0.

Sociodemographic and Lifestyle Characteristics

In addition to the physical activity status, the study's questionnaire included demographic characteristics such as age, gender, family status (married, divorced, widowed), financial status (average annual income during the past 3 yr), occupational status, and education level. The educational level of the participants (as a proxy of social status) was measured in years of school. Occupation was also recorded and was evaluated through a 10-point scale from unskilled (hand workers; lower values) to executive (skilled workers; higher values).

Information about smoking habits was collected using a standardized questionnaire developed for the study. Current smokers were defined as those who smoked at least one cigarette per day, and former smokers were defined as those who had stopped smoking for at least 1 yr. All other individuals were classified as nonsmokers. In all multivariate statistical analyses, cigarette-smoking habits were taken into account using the unit of pack-years (cigarette packs per day × years of smoking). However, to correct for the amount of nicotine (i.e., light, heavy, very heavy) contained in various types of cigarettes, we assigned a weight to each different type of cigarette pack, using 0.8 mg per cigarette as the standard.

Consumption of unrefined cereals and products, vegetables, legumes, fruits, olive oil, dairy products, fish, nuts, potatoes, eggs, sweets, poultry, red meat, and meat products were measured as average amounts consumed per week during the past year, using a validated food-frequency questionnaire (FFQ) from the department of nutritional epidemiology of Athens Medical School (10). The frequency of consumption was then quantified approximately in terms of the number of times per month that a food was consumed. Alcohol consumption was measured by daily ethanol intake, in wine glasses (100 mL and 12 g ethanol concentration). Based on the Mediterranean diet pyramid (27), we calculated a special diet score ranging from 0 to 55 (16). Higher values of this score indicate adherence to the traditional Mediterranean diet, characterized by moderate consumption of fat and a high monounsaturated:saturated fat ratio, whereas lower values indicate adherence to the "Westernized" diet.

Anthropometrics, Clinical, and Biochemical Characteristics

Standing height and weight were recorded, and body mass index (BMI) was calculated as weight (kg) divided by standing height (m2). According to standard guidelines, obesity or overweight was defined as BMI equal or greater than 25.0 kg·m−2. Waist circumference was measured on bare skin during midrespiration between the 10th rib and the iliac crest. Hip girth was measured at the maximal circumference of the buttocks. Waist-to-hip ratio (WHR) was calculated as a surrogate measure of visceral adiposity. Arterial blood pressure was measured three times, at the end of the physical examination with the subject in a sitting position. All participants had rested for at least 30 min. Patients whose average blood pressure levels were greater or equal to 140/90 mm Hg or who were under antihypertensive medication were classified as hypertensives. Blood samples were collected from an antecubital vein between 8 and 10 a.m. with the subjects in a sitting position after 12 h of fasting and avoiding alcohol. Blood glucose levels were measured immediately after collection with a Beckman glucose analyzer (Beckman Instruments, Fullerton, CA). Serum insulin concentrations were assayed by means of radioimmunoassay (RIA100, Pharmacia Co., Erlangen, Germany). Precision was 12% for low (3 μU·mL−1) and 5% for high (90 μU·mL−1) serum levels. The intraassay coefficient of variation was 9%, and the limit of detection was 3 μU·mL−1. Insulin resistance was assessed by the calculation of a homeostasis model assessment (HOMA) approach (glucose × insulin/22.5) (13). Total cholesterol was measured using a chromatographic enzymic method (Technicon automatic analyzer RA 1000, Tarrytown, NY). The intra- and interassay coefficients of variation of total cholesterol did not exceed 4%. Hypercholesterolemia was defined as a total serum cholesterol level greater than 200 mg·dL−1 or the use of lipid-lowering agents.

Statistical Analysis

Continuous variables are presented as mean values ± standard deviation, and categorical variables are presented as absolute and relative frequencies. Associations between categorical variables were tested by the use of contingency tables and chi-square test. Correlations (or partial correlations) between continuous or discrete variables were tested by the calculation of Pearson's r or Spearman's rho (for skewed variables) correlation coefficients. Comparisons between normally distributed continuous variables and categorical variables were performed by the use of analysis of covariance (multiway ANCOVA), after controlling for homoscedacity and potential confounders. The Kolmogorov-Smirnov criterion was used for the assessment of normality. In the case of asymmetric continuous variables, the tested hypotheses were based on the calculation of the Kruskal-Wallis test. Associations between the investigated variables and physical activity levels were tested through fixed-effects models after adjustment for several potential confounders and interactions with physical activity status. The assumptions of linearity and homoscedacity were graphically tested by plotting fitted values against standardized residuals. Multiple logistic regression analysis estimated the odds ratio of being at a specific health condition with respect to physical activity status. Deviance residuals evaluated the models' goodness of fit. All reported P values are based on two-sided tests and are compared using a significance level of 5%. However, because of multiple significance comparisons, we used the Bonferroni correction to account for the increase in type I error. SPSS 11.0.5 software was used for all statistical calculations.


Table 1 presents the characteristics of the participants according to their physical activity status. Five hundred sixty-five (37.3%) men and 493 (32.3%) women were classified as physically active. Men were more likely to be physically active than women across all age groups (P = 0.001). From the 1058 (34.8%) subjects who were classified as active, 306 (10.1%) met the criteria for HEPA active, and the rest were minimally active. HEPA active and minimally active subjects smoked less and had lower BMI, waist, and waist-to-hip ratio.

Participants' characteristics.

Physical Activity and Glycemic Control

Figure 1 presents the plasma glucose (A), insulin (B), and HOMA index of insulin resistance (C) presented by physical activity level and by BMI. The data show that lean and overweight or obese subjects with sedentary lifestyles had greater levels of glucose, insulin, and HOMA than active subjects. However, volunteers who were classified as overweight or obese with physical activity levels classified as HEPA had similar levels of glucose and HOMA with lean inactive individuals. Interestingly, overweight or obese subjects with HEPA active levels of physical activity had lower plasma insulin levels than lean, inactive individuals (P < 0.05).

Plasma glucose (A), insulin (B), and HOMA (C) by level of physical activity and BMI. * Statistically significant difference compared with inactive in each category of BMI (P < 0.05).

Table 2 presents the results of linear regression analysis that evaluated the association between HOMA and physical activity status, taking into consideration several social and biological factors. The results show that physical activity status (MET·min·wk−1), age, BMI, total energy intake, and presence or absence of diabetes are important predictors of HOMA, whereas other factors such as waist circumference, smoking, education, percentage of dietary energy intake as fat, hypertension, and hypercholesterolemia did not reach statistical significance.

Results from linear regression analysis that evaluated the association between homeostatic model assessment and physical activity status.


Inactivity and overweight are common characteristics in Western-type societies. In this observational study of 3042 adult people from the Attica region in Greece, we found that both physical activity and adiposity have a significant, independent effect on insulin sensitivity as assessed by the HOMA index. In addition, by examining indices of glycemic control by levels of physical activity and BMI, we have come to some remarkable results.

The major finding of these comparisons is that increased levels of physical activity may ameliorate, at least in part, the detrimental effects of overweight and obesity on insulin sensitivity. Participants with BMI ≥ 25 and physical activity levels compatible with the HEPA active category had similar levels of fasting glucose and insulin resistance, with lower levels of plasma insulin than those with normal BMI but low levels of physical activity. Thus, overweight or obese individuals may benefit from increased levels of physical activity and may show values of health indices similar to those of individuals of normal body weight. The "fit and fat" hypothesis, as a model of overall health, despite increased adiposity, originated from the Aerobics Center Longitudinal Study at the Cooper Institute in Dallas (25). In a 24-yr observational study involving nearly 26,000 men followed for an average of 10 yr, it was found that obese men with high cardiorespiratory fitness had overall mortality and cardiovascular mortality rates nearly half those of normal-weight men who were unfit. Our data are in line with these findings. Although insulin resistance per se is not considered a risk factor for mortality, it is well documented that it contributes significantly to the development of diabetes and cardiovascular disease (18,20). In addition, we have evaluated physical activity levels and not cardiorespiratory fitness, which also has a genetic component (28) and which, therefore, cannot be fully attributed to physical activity.

It should be noted that participants in the present study who belonged to the HEPA active group had large differences in insulin resistance according to BMI categorization. Lean HEPA active individuals had significantly lower levels of insulin resistance compared with overweight/obese HEPA active or lean, inactive individuals. The same trends were observed for fasting blood glucose and insulin. Two major points can be drawn from this observation. The first is that the metabolic complications (i.e., insulin resistance) of obesity and overweight cannot be fully eliminated by increased levels of physical activity. Highly active obese or overweight individuals should be encouraged to lose body weight to fully enjoy the benefits of physical activity on health. The second point is that lean, inactive individuals may develop a phenotype similar to those observed in obesity or overweight regarding insulin sensitivity. The term metabolically obese, normal-weight individual was introduced by Ruderman and coworkers in the early 1980s to describe a subgroup of the general population who display a cluster of obesity-related features, including hyperinsulinemia and dyslipidemia (19). Short-range observational studies have shown that women exhibiting this phenotype are characterized by lower levels of physical activity and cardiorespiratory fitness and higher levels of sedentary activity compared with their nonmetabolically obese counterparts (3,5). The data from the present study argue in favor of a possible causal relationship between this metabolically obese phenotype and physical activity at the population level.

Although there have been no randomized controlled clinical trials evaluating physical inactivity as the cause of obesity in a large population, there are significant amounts of observational data to support a causative link (8). In line with these studies, in the present study, HEPA active individuals had lower values of BMI and waist circumference or waist-to-hip ratio. Correspondingly, the prevalence of obesity was much lower in this activity category-less than half compared with the inactive group. Similarly, BMI but not waist circumference was a significant predictor of insulin resistance in our multivariate regression model. Although upper-body obesity has been associated with insulin resistance (8) and increased risk of cardiovascular diseases (8), a causal relationship between visceral adiposity and insulin resistance is lacking (14). Thus, overall adiposity may be a more important factor of insulin resistance.

Our study has several strengths. The study population included men as well as women, from urban as well as rural areas. It also was based on the city-gender-age distribution, allowing a representative sample of the most densely populated area of Greece to be obtained. Another major advantage of the present study is the method we used to assess physical activity levels. There are many different ways to analyze physical activity data, but, to date, there is no consensus on a criterion standard for defining levels of activity or quantifying activity on the basis of self-report surveys. We employed a questionnaire that has been validated in many populations and that could provide both qualitative (i.e., physical activity levels) and quantitative data (MET-min) on total physical activity, irrespective of the intensity of activity. Physical activity categorization is based on established recommendations for physical activity and health promotion, not on arbitrary cutoff points, thus reducing the risk of misclassification or over- and underestimation.

However, it should be pointed out that this is a cross-sectional study that cannot provide causal relationships; it can only state hypotheses for future research. Misreporting of physical activity status because of self-reporting may confound, at least in part, the strength of the observed relationships. Finally, although we took into account dietary and smoking habits of the participants, the influence of the potential confounding effects of these factors cannot be excluded entirely.

In conclusion, insulin resistance is strongly associated with physical activity status, independently of the presence or absence of obesity. Given the increasing prevalence of diabetes and other metabolic diseases, physical activity promotion should be the cornerstone of public health policies for the whole population. In addition, clinicians should encourage patients or those at risk of developing metabolic diseases to increase their physical activity levels.

The ATTICA Study is funded by research grants from the Hellenic Society of Cardiology (grant 1, 2002).

The authors would like to thank the field investigators of the ATTICA study: John Skoumas (principal field investigator), Natasa Katinioti (physical examination), Spiros Vellas (physical examination), Efi Tsetsekou (physical/psychological evaluation), Dina Masoura (physical examination), Lambros Papadimitriou (physical examination); and the technical team: Marina Toutouza (biochemical analysis), Carmen Vasiliadou (genetic analysis), Manolis Kambaxis (nutritional evaluation), Konstadina Paliou (nutritional evaluation), Chrysoula Tselika (biochemical evaluation), Sia Poulopoulou (biochemical evaluation), and Maria Toutouza (database management).


1. Ainsworth, B. E., W. L. Haskell, M. C. Whitt, et al. Compendium of physical activities: an update of activity codes and MET intensities. Med. Sci. Sports Exerc. 32:S498-S504, 2000.
2. Carlsson, S., K. Midthjell, M. Y. Tesfamarian, and V. Grill. Age, overweight and physical inactivity increase the risk of latent autoimmune diabetes in adults: results from the Nord-Trondelag health study. Diabetologia 50:55-58, 2007.
3. Conus, F., D. B. Allison, R. Rabasa-Lhoret, et al. Metabolic and behavioral characteristics of metabolically obese but normal-weight women. J. Clin. Endocrinol. Metab. 89:5013-5020, 2004.
4. Craig, C. L., A. L. Marshall, M. Sjostrom, et al. International physical activity questionnaire: 12-country reliability and validity. Med. Sci. Sports Exerc. 35:1381-1395, 2003.
5. Dvorak, R. V., W. F. DeNino, P. A. Ades, and E. T. Poehlman. Phenotypic characteristics associated with insulin resistance in metabolically obese but normal-weight young women. Diabetes 48:2210-2214, 1999.
6. Eriksson, K. F., and F. Lindgarde. Prevention of type 2 (non-insulin-dependent) diabetes mellitus by diet and physical exercise. The 6-year Malmo feasibility study. Diabetologia 34:891-898, 1991.
7. Flegal, K. M., M. D. Carroll, C. L. Ogden, and C. L. Johnson. Prevalence and trends in obesity among US adults, 1999-2000. JAMA 288:1723-1727, 2002.
8. Grundy, S. M., G. Blackburn, M. Higgins, R. Lauer, M. G. Perri, and D. Ryan. Physical activity in the prevention and treatment of obesity and its comorbidities: evidence report of independent panel to assess the role of physical activity in the treatment of obesity and its comorbidities. Med. Sci. Sports Exerc. 31:1493-1500, 1999.
9. Hu, F. B., R. J. Sigal, J. W. Rich-Edwards, et al. Walking compared with vigorous physical activity and risk of type 2 diabetes in women: a prospective study. JAMA 282:1433-1439, 1999.
10. Katsouyanni, K., E. B. Rimm, C. Gnardellis, D. Trichopoulos, E. Polychronopoulos, and A. Trichopoulou. Reproducibility and relative validity of an extensive semi-quantitative food frequency questionnaire using dietary records and biochemical markers among Greek schoolteachers. Int. J. Epidemiol. 26:S118-S127, 1997.
11. Knowler, W. C., E. Barrett-Connor, S. E. Fowler, et al. Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. N. Engl. J. Med. 346:393-403, 2002.
12. Lillioja, S., D. M. Mott, M. Spraul, et al. Insulin resistance and insulin secretory dysfunction as precursors of non-insulin-dependent diabetes mellitus. Prospective studies of Pima Indians. N. Engl. J. Med. 329:1988-1992, 1993.
13. Matthews, D. R., J. P. Hosker, A. S. Rudenski, B. A. Naylor, D.F. Treacher, and R. C. Turner. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 28:412-419, 1985.
14. Miles, J. M., and M. D. Jensen. Counterpoint: visceral adiposity is not causally related to insulin resistance. Diabetes Care 28: 2326-2328, 2005.
15. Pan, X. R., G. W. Li, Y. H. Hu, et al. Effects of diet and exercise in preventing NIDDM in people with impaired glucose tolerance. The Da Qing IGT and Diabetes Study. Diabetes Care 20:537-544, 1997.
16. Panagiotakos, D. B., C. Pitsavos, C. Chrysohoou, J. Skoumas, and C. Stefanadis. Status and management of blood lipids in Greek adults and their relation to socio-demographic, lifestyle and dietary factors: the ATTICA Study. Blood lipids distribution in Greece. Atherosclerosis 173:353-361, 2004.
17. Pate, R. R., M. Pratt, S. N. Blair, et al. Physical activity and public health. A recommendation from the Centers for Disease Control and Prevention and the American College of Sports Medicine. JAMA 273:402-407, 1995.
18. Reaven, G. M., H. Lithell, and L. Landsberg. Hypertension and associated metabolic abnormalities-the role of insulin resistance and the sympathoadrenal system. N. Engl. J. Med. 334:374-381, 1996.
19. Ruderman, N. B., S. H. Schneider, and P. Berchtold. The "metabolically-obese," normal-weight individual. Am. J. Clin. Nutr. 34:1617-1621, 1981.
20. Steinberg, H. O., H. Chaker, R. Leaming, A. Johnson, G. Brechtel, and A. D. Baron. Obesity/insulin resistance is associated with endothelial dysfunction. Implications for the syndrome of insulin resistance. J. Clin. Invest. 97:2601-2610, 1996.
21. Tanasescu, M., M. F. Leitzmann, E. B. Rimm, W. C. Willett, M.J. Stampfer, and F. B. Hu. Exercise type and intensity in relation to coronary heart disease in men. JAMA 288:1994-2000, 2002.
22. Tuomilehto, J., J. Lindstrom, J. G. Eriksson, et al. Prevention of type 2 diabetes mellitus by changes in lifestyle among subjects with impaired glucose tolerance. N. Engl. J. Med. 344:1343-1350, 2001.
23. Wannamethee, S. G., and A. G. Shaper. Weight change and duration of overweight and obesity in the incidence of type 2 diabetes. Diabetes Care 22:1266-1272, 1999.
24. Wannamethee, S. G., A. G. Shaper, and K. G. Alberti. Physical activity, metabolic factors, and the incidence of coronary heart disease and type 2 diabetes. Arch. Intern. Med. 160:2108-2116, 2000.
25. Wei, M., J. B. Kampert, C. E. Barlow, et al. Relationship between low cardiorespiratory fitness and mortality in normal-weight, overweight, and obese men. JAMA 282:1547-1553, 1999.
26. Wild, S., G. Roglic, A. Green, R. Sicree, and H. King. Global prevalence of diabetes: estimates for the year 2000 and projections for 2030. Diabetes Care 27:1047-1053, 2004.
27. Willett, W. C., F. Sacks, A. Trichopoulou, et al. Mediterranean diet pyramid: a cultural model for healthy eating. Am. J. Clin. Nutr. 61:1402S-1406S, 1995.
28. Wolfarth, B., M. S. Bray, J. M. Hagberg, et al. The human gene map for performance and health-related fitness phenotypes: the 2004 update. Med. Sci. Sports Exerc. 37:881-903, 2005.


©2007The American College of Sports Medicine