BACKGROUND: A global epidemic of obesity is developing. Current prevalence rates are about 20–25% in American adults and 15–20% in Europeans.
OBJECTIVE: We investigated the association between population levels of physical activity, dietary fat, dietary fiber and indicators of body fat.
DESIGN: Cross-cultural study of 16 cohorts of, in total, 12 763 middle-aged men in seven countries. These men were examined between 1958 and 1964.
MEASUREMENTS: Height, weight and subscapular skinfold thickness were measured. Information about job-related physical activity and diet was gathered by questionnaire.
RESULTS: The population average body mass index (weight/height2) varied between 21.8 and 26.0 kg/m2 and the population average subscapular skinfold thickness between 8.4 and 23.7 mm. The population average physical activity index and dietary fiber intake were both strongly inversely related to population average subscapular skinfold thickness and explained together 90% of the variance in subscapular skinfold thickness. Similar but less strong results were obtained for average population body mass index.
CONCLUSION: At the population level job-related physical activity and dietary fiber but not dietary fat, are important determinants of subscapular skinfold thickness.
Obesity is a common and growing problem in affluent countries. Current prevalence rates are about 20–25% in American adults and 15–20% in Europeans.1 Direct costs of obesity in these regions have been estimated to be responsible for about 5–10% of total health care costs and indirect costs (due to loss of productivity) are of the same order of magnitude.2,3 The problem of obesity is smaller in many countries undergoing economic transition, but is rapidly increasing there. This increase is accompanied by a sharp increase in the incidence of type 2 diabetes mellitus and cardiovascular diseases.4
The development of obesity is the consequence of chronic and sustained energy imbalance when energy intake exceeds energy expenditure. Therefore it is necessary to take both sides of the energy equation simultaneously into account. Both energy imbalance as well as maintenance of obesity are thought to be facilitated by energy-dense diets and sedentary lifestyles.3 Short-term covert manipulation of energy density of food has shown to affect energy intake.5,6
The role of fat intake in the etiology of obesity is controversial. Epidemiological and long-term intervention studies did show a relatively small effect of changes in dietary fat content on energy balance and weight change.7,8 Ecological studies on the association between relative fat intake and body mass index (BMI, weight/height2) have shown both positive and inverse associations.9 These inconsistencies can be partly explained by the lack of control for energy expenditure and energy-dense diets. Data from the UK have suggested that crude indicators of leisure time physical activity such as amount of television viewing and number of cars in the household are strongly associated with trends in obesity rates and differences between socio-economic classes, whereas the contribution of energy intake by dietary fat was not related to obesity.10
To our knowledge there have been no attempts to study differences in obesity rates in populations in relation to the combination of population energy expenditure levels and indicators of energy density. In addition, relative fat intake was usually the sole indicator of dietary intake and energy expenditure was evaluated mainly by leisure time activities. Moreover, in virtually all studies BMI was used as a measure of body fat and the validity of BMI as a measure of body fat across populations has been questioned.11,12
In the present study we used both average levels of body fat measured by BMI and subscapular skinfold thickness in middle-aged men participating in the Seven Countries Study in the period 1958–1964. The study populations ranged from farmers to university professors and differences in energy expenditure were predominantly caused by differences in job-related physical activity and not by leisure time activity. Indicators of energy density included not only fat intake but also fiber intake.
Between 1958 and 1964, 12 763 men aged 40–59 were enrolled in the Seven Countries Study. In these countries 16 cohorts were established, 11 in rural areas in Finland, Italy, Greece, the former Yugoslavia and Japan, two cohorts of railroad employees in the USA and Italy, one of workers in a large co-operative in Serbia, one of university professors in Belgrade and one of inhabitants of a small commercial market town in The Netherlands. The characteristics of these cohorts have been described in detail.13,14,15
All men in the Seven Countries Study were classified at entry according to their job-related habitual physical activity pattern.13 Class 1 men were bedridden. Class 2 men were sedentary and engaged in little exercise. This category embraced clerks, salesmen, managers, teachers etc. Class 3 men were moderately active during a substantial part of the day and consisted of bakers, butchers, plumbers, small shop-keepers etc. Class 4 men performed hard physical work much of the time and examples of this category are farmers, lumberjacks, fishermen, blacksmiths etc. For each cohort the percentage of men in each class was determined and the average calculated. The average index was multiplied by 1000 in order to get values in the order of magnitude comparable to energy expenditure expressed in kcal.
Weight was measured with a balance to the nearest 100 g without shoes, in light undergarment, according to a standardized protocol.13 Height was measured to the nearest 0.5 cm without shoes. The subscapular skinfold thickness was measured twice in a standardized way at the right side of the body. The average was used in the analysis. For the Japanese cohorts Tanushimaru and Ushibuka the subscapular skinfold thickness values of the 10 y follow-up survey were used because this skinfold was not measured at the baseline examination.
Between 1959 and 1964 dietary information was collected in small random samples of 14 of the 16 cohorts.16 In the other two cohorts dietary information was gathered around 1970. The weighed record method was used in all dietary surveys. In 1985 and 1986 the original dietary data of these cohorts were coded by one dietitian in a standardized way. The average intake of different foods consumed in the 16 cohorts was calculated and summarized in 16 food groups.
In 1987, equivalent food composites representing the average food intake of each cohort at baseline were collected from local markets by two dietitians. These foods were transported in cooling boxes to the laboratory of the Department of Human Nutrition, Agricultural University, Wageningen, The Netherlands (Head, MB Katan). The foods were cleaned and equivalent food composites prepared according to the average consumption patterns of the cohorts. The foods were homogenized and frozen at −20°C until chemical analyses of the different nutrients took place. Total lipids were isolated according to Osborne and Voogt.17 Total energy intake is the sum of total protein, fat and carbohydrates. Total dietary fiber was determined with an enzymatic gravimetric method.18
The physical activity index, dietary intake variables, BMI and subscapular skinfold thickness represent the average for each cohort. Analyses concern only inter-cohort comparisons. Because of the limited degrees of freedom, the multiple regression models never included more than four independent variables. Only two-sided P-values are reported.
The population average BMI varied between 21.8 in Tanushimaru (Japan) and 26.6 kg/m2 in the Rome Railroad (Italy) (Table 1). The population average subscapular skinfold thickness varied between 8.4 in Velika Krsna (Serbia) and 23.7 mm in Belgrade (Serbia). The correlation coefficient between the population average BMI and subscapular skinfold thickness was 0.75 (P<0.001). The average BMI of all cohorts increased from 24.0 to 24.8 kg/m2 between 1960 and 1970. The correlation coefficient between the population average BMI values measured 10 y apart was 0.93 (P<0.001). The average subscapular skinfold thickness of all cohorts increased between 1960 and 1970 from 12.5 to 14.8 mm. The correlation between the population average subscapular skinfold thickness measured 10 y apart was 0.89 (P<0.001).
The population average physical activity index varied from 2006 in Belgrade (Serbia) to 3982 in Velika Krsna (Serbia). Only 13 of the 12 763 men were bedridden. The population average energy intake varied from 2251 kcal/day in the American Railroad to 3609 kcal/day in Slavonia (Croatia). The correlation between the population average physical activity index and energy intake was 0.45 (P<0.10). Dietary fat intake varied from 33 g/day in Tanushimaru (Japan) to 179 g/day in Slavonia (Croatia). Men in Ushibuka (Japan) consumed 21 g of fiber per day compared with 57 g/day in men from Corfu (Greece).
The population average physical activity index was significantly inversely associated with population average BMI (Figure 1). The population average physical activity index was also significantly inversely associated with population average BMI after multivariate analyses (Table 2). Population average BMI was positively associated with population average fat intake and inversely with dietary fiber intake. These associations did not reach statistical significance after both uni and multivariate analyses.
Population average physical activity index was inversely associated with population average subscapular skinfold thickness (Figure 2). In both uni and multivariate analyses population average fat intake was not associated with population average subscapular skinfold thickness (Table 3). Population average dietary fiber intake, however, was significantly inversely associated with population average subscapular skinfold thickness in both uni- and multivariate analyses (Figure 3 and Table 3). Population average physical activity index and dietary fiber together explained 90% of the variance in population average subscapular skinfold thickness.
The results of this ecological analysis show that physical activity and dietary fiber intake are both strong and independent inverse correlates of average subscapular skinfold thickness in middle-aged men living in the 1960s under very different cultural circumstances. Dietary fat intake was not a significant correlate of average subscapular skinfold thickness in the present study. Results were generally in the same direction but less strong for average BMI. This probably reflects the limitations to the use of BMI as a measure of body fat. These results suggest that physical activity and dietary fiber are the most important environmental determinants of population levels of body fat.
The weakness of the present study is that this analysis is ecological and cross-sectional and covers 16 cohorts only. There are, however, also many merits to this study. One of the strengths of the study is the variation in average levels of body fat in the populations studied (range of average subscapular skinfold thickness, 8–24 mm, and range for BMI, 22–27 kg/m2). The data were collected at a time when most of the variation in total daily energy expenditure in middle-aged men was determined by the level of physical activity at work. In addition, this is to our knowledge the first study to investigate cross-cultural differences in levels of body fat examining key indicators of energy intake, energy density and energy expenditure in one single data-set. The timing of data collection in the Seven Countries Study limits the extrapolation to the contemporary situation but the results are probably applicable for many countries in the world in different phases of economic transition. Another advantage is the availability of subscapular skinfold thickness in addition to BMI.
For ecological analyses it is important to know how stable the position is of each cohort in the distribution of a body fat indicator. Although the population average BMI of 16 cohorts increased from 24.0 to 24.8 kg/m2 and the population average subscapular skinfold thickness from 12.5 to 14.8 mm during a follow-up period of 10 y, the correlation coefficients for these indicators of body fat measured 10 y apart amounted to about 0.9 (P<0.001). This means that, in spite the increase in the population average values of these indicators of body fat, their relative position in distributions of these indicators remained the same. Therefore these populations were very well suited to study the associations between physical activity, dietary variables and body fat indicators at the population level.
In epidemiological studies BMI is often used as indicators for body fat. BMI is often assumed to be independent of height (although it is not19) and correlates strongly with body fat. However a high BMI may be due to a large amount of fat but can also be due to a large amount of muscle or bone. Therefore the correlation between BMI and body fat is not perfect. Subscapular skinfold thickness is a better and more direct measure of body fat than BMI. In the present study the correlation between the population average BMI and subscapular skinfold thickness was 0.75 (P<0.001). This indicates that, although these two indicators of body fat are strongly correlated, they do not exactly measure the same. In terms of indicators of body fat subscapular skinfold thickness is superior to BMI. Subscapular skinfold thickness moreover reflects truncal fat and has been found to be a stronger and independent predictor of coronary heart disease than BMI.20,21
The present study shows that population average physical activity is inversely related to both population average BMI and subscapular skinfold thickness. The association was, however, stronger for subscapular skinfold thickness than for BMI. These results are in accord with those from a study on secular trends carried out in Britain.10 The secular trend on the prevalence of obesity between 1950 and 1990 was strongly related to that of the number of hours per week of television viewing, an indicator of physical inactivity. Also these results suggest that at the population level physical activity is a strong determinant of body fat. The patterns of physical activity explaining recent trends in the UK have of course shifted from physical activity at work to variation in leisure time activity since most jobs now tend to be more or less sedentary.
The energy density of the diet is generally viewed as the main dietary determinant of body fat.3 Indeed, covert and overt manipulation of energy density of diets leads to spontaneous changes in total energy intake under ad-libitum conditions.5,6 Although diets with a high fat content usually have a high energy density, fat intake is not the sole determinant of energy density of the diet; fiber intake is another important one.6
The relationship between dietary fat intake and body fat is controversial. This is partly due to the fact that most epidemiological studies have reported only associations between percentage energy derived from fat in the diet and body fat. Those studies failed to adjust for the absolute intake of fat and fiber and indicators of energy expenditure.22 In the present study significant associations between dietary fat intake and either BMI or subscapular skinfold thickness were not observed. This implies that in these populations, compared to a low-fat diet, a high dietary fiber intake was more important for the protection against high levels of body fat. High fiber intake leads to ‘bulky’ diets which limit spontaneous intake of energy.5,6
When the Seven Countries Study started, physical activity was largely determined by physical activity on the job. Recreational physical activity did not play an important role in the mostly rural populations enrolled in the Seven Countries Study. Most of the jobs in the cohorts with low levels of body fat were physically very demanding, like farming, with little mechanical support. Such jobs are rare in most established market economies but are still very common in other parts of the world. Under the influence of the economies transition in Asia, Latin America and Africa, physical activity during work is rapidly diminishing, particularly in urban areas.23 At the same time dietary changes in these areas lead to an increasing fat and decreasing fiber intake. The results of the present study indicate that, in particular, the reduced job-related energy expenditure and the reduced fiber intake may be strong determinants of increasing body fat levels in these populations.
Migration and time-trend studies have suggested that, at the population level, genetic factors play a minor role in explaining differences in body fat level.24 This can be illustrated by comparing the results of two Serbian cohorts in the Seven Countries Study. The farmers in Velika Krsna and the Belgrade professors have a similar ethnic background. At the same time they represent the extremes in subscapular skinfold thickness distribution within the Seven Countries Study (Table 1). The average subscapular skinfold thickness of the farmers in Velika Krsna is 8 mm and the average in Belgrade professors 24 mm. Around 1960 the farmers had a high job-related level of physical activity and a relatively high fiber intake. The professors were physically inactive and had a low fiber intake. These factors largely determined their level of a body fat.
The present study has shown that there are two important determinants of body fat at the population level: physical activity and a high-fiber diet. Increasing levels of body fat in the populations can be regarded as a normal response to changes in these life-style factors,25 although individual variation in weight gain in response to a high-fat diet and low levels of physical activity may be determined by genetic susceptibility.26,27 Both a low level of physical activity and a low-fiber diet are not only determinants of body fat but also of several chronic diseases such as coronary heart disease and colon cancer. If more evidence is accumulating on the causal relation between these two factors and body fat, demonstration programs can be designed to show that an increase in physical activity and dietary fiber can prevent obesity.
Seidell, JC, Flegal KM . Assessing obesity: classification and epidemiology Br Med Bull 1997 53: 238–252.
Seidell JC . The impact of obesity on health status: some implications for health care costs Int J Obes Relat Metab Disord 1995 19 (Suppl 6): S13–S16.
World Health Organization . Obesity: preventing and managing the global epidemic World Health Organization: Geneva 1998.
Amos AF, McCarty DJ, Zimmet P . The rising global burden of diabetes and its complications: estimates and projections to the year 2010 Diabet Med 1997 14 (Suppl 5): S1–S85.
Prentice AM . Manipulation of dietary fat and energy density and subsequent effects on substrate flux and food intake Am J Clin Nutr 1998 67: 535S–541S.
Rolls BJ, Bell EA, Castellanos VH, Chow M, Pelkman CL, Thorwart ML . Energy density but not fat content of foods affected energy intake in lean and obese women Am J Clin Nutr 1999 69: 863–871.
Bray GA, Popkin BM . Dietary fat intake does affect obesity! Am J Clin Nutr 1998 68: 1157–1173.
Willett WC . Dietary fat and obesity: an unconvincing relation Am J Clin Nutr 1998 68: 1149–1150.
Lissner L, Heitmann BL . Dietary fat and obesity: evidence from epidemiology Eur J Clin Nutr 1995 49: 79–90.
Prentice AM, Jebb S . Obesity in Britain: gluttony or sloth? Br Med J 1995 311: 437–439.
Deurenberg P, Yap M, van Staveren WA . Body mass index and percent body fat: a meta analysis among different ethnic groups Int J Obes Relat Metab Disord 1998 22: 1164–1171.
Heitmann BL, Swinburn B, Carmichael H, Rowley K, Plank L, McDermott R, Leonard D, O'Dea K . Are there ethnic differences in the association between body weight and resistance, measured by bioelectrical impedance? Int J Obes Relat Metab Disord 1997 21: 1085–1092.
Keys A, Aravanis C, Blackburn HW et al. Epidemiological studies related to coronary heart disease: characteristics of men aged 40–59 in seven countries Acta Med Scand 1967 460: 1–392.
Keys A (ed) . Coronary heart disease in seven countries Circulation 1970 41 (Suppl 1): 1–211.
Keys A . Seven countries: a multivariate analysis of death and coronary heart disease Harvard University Press: Cambridge, MA 1980.
Kromhout D, Keys A, Aravanis C et al. Food consumption patterns in the 1960s in seven countries Am J Clin Nutr 1989 49: 889–894.
Osborne DR, Voogt P . Soxhlet method In: Analysis of nutrients and foods 1st edn: Academic Press: London 1978 155–156.
William S (ed) . Total dietary fiber in foods; enzymatic gravimetric method. Changes in methods J Ass Off Anal Chem 1985 68: 399.
Han TS, Seidell JC, Curral JEP, Morrison CE, Deurenberg P, Lean MEJ . The influence of height on waist circumference as an index of adiposity in adults Int J Obes Relat Metab Disord 1997 21: 83–89.
Donahue RP, Abbott RD, Bloom E, Reed DM, Yano K . Central obesity and coronary heart disease in men Lancet 1987 i: 821–824.
Higgins M, Kannel WB, Garrison R, Pinski J, Stokes J III . Hazards of obesity—the Framingham experience Acta Med Scand 1988 723 (Suppl): 23–36.
Seidell JC . Dietary fat and obesity: an epidemiologic perspective Am J Clin Nutr 1998 67 (Suppl): 546S–550S.
Drenowski A, Popkin BM . The nutrition transition: new trends in the global diet Nutr Rev 1997 55: 31–43.
Hill JO, Peters JC . Environmental contributions to the obesity epidemic Science 1998 280: 1371–1374.
Egger G, Swinburn B . An ‘ecological’ approach to the obesity pandemic Br Med J 1997 315: 477–480.
Heitmann BL, Lissner L, Sörensen TIA, Bengtsson C . Dietary fat intake and weight gain in women genetically predisposed for obesity Am J Clin Nutr 1995 61: 1213–1217.
Heitmann BL, Kaprio J, Harris JR, Rissanen A, Korkeila M, Koskenvuo M . Are genetic determinants of weight gain modified by leisure-time physical activity? A prospective study of Finnish twins Am J Clin Nutr 1997 66: 672–678.
The authors are grateful to the principal investigators in the different countries who initiated the Seven Countries Study and for their effort in carrying out the study for more than 25 y.
About this article
Assessment of total fatness and fatty tissue distribution in young active and physically inactive women
Biomedical Human Kinetics (2018)
Changes in 10-year cardiovascular risk and behavioral risk factors in men in Crete, Greece, since the Seven Countries’ Study (1960–1997)
Journal of Public Health (2018)
Diet quality as measured by the Diet Quality Index–International is associated with prospective changes in body fat among Canadian children
Public Health Nutrition (2017)
Taiikugaku kenkyu (Japan Journal of Physical Education, Health and Sport Sciences) (2014)