Fat free mass and obesity in relation to educational level

Obesity is characterized by excess adipose tissue. Quantification of adipose tissue mass can be achieved by a number of laboratory methods including underwater body density measurement, dual energy X-ray absorptiometer (DEXA) and magnetic resonance imaging [1‐3]. These methods require costly equipment and are difficult to implement in epidemiological studies. Some exceptions, such as bioelectrical impedance (BIA) exist[4]. To date, however, population surveys using BIA are scarce. Body weight adjusted for stature (body mass index) is yet commonly used as a surrogate for body fat content in population surveys[5, 6].

A strong inverse association between overall obesity, as defined by body mass index (BMI), and socio-economic status, mostly assessed by educational level, is well defined in affluent populations [7‐14]. In general, this association does not seem to be as consistent for men as for women [10, 12, 14, 15].

To date, obesity measurements other than BMI have seldom been used to examine the relationship between socio-economic status and obesity. In a few studies waist circumference or waist-to-hip ratio (WHR), as measures of abdominal obesity, has been shown to be inversely associated with education both for men and women [14, 16‐21] or only for men[22].

Because BMI cannot distinguish between fat mass and fat-free mass (FFM), it is not known from which component the educational gradient of BMI is composed of. Although FFM has been shown to be positively associated with BMI [23] and also weakly associated with WHR, [24]. it remains unknown, whether the educational differences are due to differences in muscularity or in fatness.

Obviously, people carrying out more physical activity have higher FFM than less active people. To our knowledge, however, there are no studies on the association between FFM and education. It has been observed that high socio-economic status (SES) or high level of education is associated with less job-related physical activity[25]. Therefore, on one hand, we hypothesized that less educated people might be more muscular and have higher FFM than people with higher education due to manual work. However, as the typical jobs today are physically much less demanding than the working conditions some decades ago [26], these possible associations might vary by age. On the other hand, education has been shown to be associated with more frequent leisure time physical activity, [25, 27] suggesting that due to being more active at leisure time subjects with high SES could have higher FFM than their counterparts in lower SES groups.

The aim of the present study was to describe body composition of Finnish adult population, especially using FFM assessed by BIA as an indicator, and its relation to educational level. Our further aim was to investigate the possible effects of the variation in FFM on the educational gradients in BMI and WHR, i.e. whether FFM can explain the educational gradients.

Weight, height, body mass index (BMI), waist-to-hip ratio (WHR) and fat free mass (FFM) of the subjects are presented by sex and age in Table 1. In men, the mean weight, BMI and WHR increase with age until the age of 65 years, whereas the mean values of height and FFM decrease constantly with age. Similarly in women, the mean weight increases with age up to 65 years of age and the mean BMI up to 75 years of age. In women, however, the higher the age is, the lower are the mean values of height and FFM, and higher of WHR.

In the total study population, the mean BMI and WHR varied across educational groups both in men (Table 2) and women (Table 3). The values were highest among the least educated subjects, and lowest among the most educated subjects. In men, the educational gradient was observed in middle-aged (30-64 y.) men only (Table 2), whereas in women, the mean BMI and WHR varied by education in both age groups (Table 3). No differences in the mean FFM were observed in relation to education in the total study population (Tables 2 and 3). Nevertheless, the relation between the mean FFM and education seemed to vary by age and sex, such that the difference between educational groups in the mean FFM was statistically significant in older men (≥ 65 y) only. In this group, the mean FFM was lowest among the least educated men (Figure 1).

The analysis of effect modification revealed that there were significant age-adjusted interactions between sex and education for low (the lowest quintile) BMI (P < 0.001) and low (the lowest quintile) FFM (P < 0.001). For the other outcomes the corresponding p-values varied between 0.15 and 0.47. In men, the interaction term between age and education was close to being statistically significant for high (the highest quintile) WHR (P = 0.08) and low FFM (P = 0.08), whereas in women, the effect modification by age was weaker. Due to these results, the final analyses for investigating associations of education with low and high body composition values were stratified into four groups: men and women aged 30-64 years or aged 65 years or older.

The age-adjusted associations between education and FFM were quite weak (Model 1, Table 4). No associations were observed in older age groups but in the middle-aged group, highly educated men were less likely to have a low FFM compared with men in the lowest education group. With adjustment for BMI, also men with middle level education and women with high level education in older groups were less likely to have a low FFM compared with their least educated counterparts (Model 2, Table 4).

With adjustment for age, in middle-aged men, those with middle or high level education were less likely to have a high BMI compared with their less educated counterparts (Model 1, Table 5). Women with the highest level of education in both age groups were less likely to have a high BMI than women with low education (Model 1, Table 5). In middle-aged men, those with middle or high level education were less likely to have a high WHR than men with the lowest education (Model 1, Table 6). In both age groups, women with the highest level of education were less likely to have a high WHR than their least educated counterparts (Model 1, Table 6). In general, adjustment for FFM seemed to strengthen the associations of education with the prevalence of high BMI or high WHR (Model 2, Tables 5 and 6). For example, in the middle-aged group, men and women with middle level education seemed to be less likely to have a high BMI compared with their less educated counterparts. In women, this association reached statistical significance after adjustment for FFM.

The association between education and FFM was fairly weak. Highly educated men, however, were less likely to have a low FFM suggesting that highly educated individuals have more muscles compared with their less educated peers. This association was not explained by differences in BMI as after adjustment for BMI, the associations remained in the middle-aged men and strengthened in older men and women. Not surprisingly, educational level was inversely associated with overall (high BMI) and abdominal (high WHR) obesity both in men and women. FFM was not found to explain these educational gradients but adjustment for FFM seemed to slightly strengthen the inverse associations between education and obesity.

To our knowledge, BIA based data on body composition in representative population sample have not earlier been reported. More importantly, to date, the present study is the first one that has investigated whether there is any relation between FFM and educational level.

Our findings that education was inversely associated with the prevalence of overall obesity both in men and women are consistent with many earlier studies both in Finland [11, 12, 14] and other countries[7‐10, 13]. In agreement with our results, the findings have been more consistent in women than in men[10, 12, 14, 15]. In addition, in most of the previous studies, central obesity has also been shown to be inversely associated with education both in men and women [14, 16‐21].

In the majority of the earlier studies concerning the association between education and obesity, the subjects have been mainly under 65 years of age. In the present study, the inverse associations between education and both overall and abdominal obesity were evident in older women but not in older men, which supports the findings of a recent study from Spain, [36] where inverse associations between education and BMI and waist circumference were evident in older (≥ 60 year) women but not in older men. Contrary to our findings, another study from Chicago [37] found BMI to be inversely associated with education both in men and women in older (≥ 64 year) groups.

The reasons for gender differences in the educational patterning of obesity by age are not known. The association between socio-economic status and obesity is complex: it is bidirectional and confounded by other factors such as heredity, [38] health behaviour [25, 39] or in general social and cultural norms[15]. It is possible that a strong desire to be thin among the highest educated women [25, 39] lasts throughout their entire lives. It is also possible that a few decades ago, obesity and especially abdominal obesity was a sign of high socio-economic status in men and was thus a more acceptable characteristic for men. This may be the reason for the diminished association between obesity and education among the older men.

In older men, the standard deviation in FFM varied across educational groups, being largest in the least educated men, and narrowest in the highly educated men. This finding indicates larger heterogeneity in body composition in the least educated older men. It is possible that due to this heterogeneity, the association between FFM and education has diminished. It is also worth of noticing that despite the lack of educational gradient in BMI for older men, the mean FFM was smaller in men with low education compared with their more educated counterparts. This finding suggests that those men may have more adverse body composition. One explanation for this might be more frequent co-morbid conditions (e.g. heart failure or chronic obstructive pulmonary disease) that have been shown to be related to lower FFM[40, 41]. Contrary to this assumption, however, the previous Finnish study showed lower socio-economic status to be associated with smaller waist circumference in working-aged men when BMI was adjusted for[17]. In all, socio-economic patterning of obesity has shown to be divergent. For example, being thin, as defined by low BMI, has been observed to be associated with unemployment and low income in men [42, 43] and with low income in women[42].

Contrary to our expectations, there was no evidence that high FFM is a characteristic of people with low educational status. Our findings did not support our hypothesis of less educated people due to manual work to have higher FFM than people with higher education. Our results do indeed suggest that less-educated middle-aged individuals have less muscle compared with their more educated peers. Compared with the working conditions of the early 1980s, the typical jobs today are physically much less demanding[26]. Physical activity at work decreased, whereas leisure-time activity increased in Finland between the 1980s and the 1990s[19, 26]. In all, during the same period, total physical activity decreased[26].

We assumed that people carrying out more physical activity have higher FFM than people carrying out less physical activity. However, the association between physical activity and FFM is not so simple. The type of exercise, aerobic or resistance exercise, may have different impacts on FFM[44]. Furthermore, physical activity has been observed to be positively [16] or inversely [45] associated with educational level in men. In Eastern Finland in the 1970s, people carrying out a lot of physical activity were better educated than people who did not carry out much physical activity.46] In the early 1980s, this educational gradient disappeared in both genders, but in 1997 men who did not carry out much physical activity were better educated than those who did. It has also been observed that high socio-economic status or high level of education is associated with less job-related physical activity [25, 47] but more leisure time fitness activity[25, 27, 47]. It may be that several decades ago, less educated, manual workers probably did have more muscles compared with higher educated people, whereas nowadays highly educated people who may be more physically active in their leisure time may be more muscular.

The variation in FFM could not explain the association of education with overall and abdominal obesity but it strengthened these educational gradients to some extent. This was somewhat unexpected, because FFM is one of two important body composition components (i.e., fat mass and FFM), which influence both BMI and WHR although it has less influence on WHR. Men generally have a higher mean FFM than women, who in turn have higher fat mass and fat percentage than men [48‐50]. Furthermore, BMI cannot distinguish between fat mass and FFM, and increased WHR may reflect decreased muscle mass in the lower part of the body as much as increased abdominal adiposity[51, 52]. Therefore we assumed that FFM could confound and explain the different association between education and obesity among men and women. Indeed, our results of FFM strengthening the associations may suggest that differences in overall obesity defined by BMI and abdominal obesity defined by WHR indicate quite well differences in fat mass.

The strengths of our study include a database with a large number of subjects and trained personnel. A further strength was that a comprehensive set of methods was used to measure body composition.

Regardless of the large number of subjects in our study, there were only a few subjects in the older age groups, especially men, in the middle and high educational level after dividing subjects into the lowest and highest quintiles of BMI, WHR and FFM. For these kinds of analyses sample size should be even larger to give the analyses more statistical power.

The field conditions of the health examination were standardized as far as possible[28, 29]. All the subjected had been asked to come to the examination after fasting on the same day. However, the examinations were started between 8 o'clock in the morning and 2 o'clock in the afternoon, sometimes even later, so that BIA results which depend on the time of day and the duration of fasting may vary because of variations in the conditions.

In the future, possible confounding factors such as physical activity, eating habits and chronic diseases should be taken into account to improve understanding of association between deviant body composition and education. In addition, prospective studies should give more information on causes and consequences of deviant body composition and education.

The inverse association of educational level with overall and abdominal obesity cannot be explained by FFM. Adjustment for FFM may even strengthen the educational gradient. Contrary to our first hypothesis, high FFM is not a characteristic of people with a low level of education.