Year: 2005
Type: review
Abstract: There is a worldwide increase in the prevalence of obesity,1 which contributes to a higher incidence of cardiovascular disease and type 2 diabetes mellitus.2,3 It has been projected that by 2020, type 2 diabetes and cardiovascular disease will account for almost three-fourth of all deaths worldwide.4 Since the 1950s it has been recognized that apart from overall obesity the distribution of body fat can influence disease risk.5 In particular an abdominal fat distribution is associated with metabolic disturbances and increased risk of cardiovascular disease and type 2 diabetes. In the present paper we will discuss different ways to measure body composition and focus on which aspects of body fatness (e.g. central vs total, subcutaneous vs visceral) are particularly hazardous in terms of morbidity and mortality. Numerous techniques are available to estimate body composition and fat distribution, and the method to use will depend on the aim of the study, economic resources, availability, time, and sample size.6–8 Multi-compartment models, such as underwater weighing, dilution techniques and dual-energy X-ray absorptiometry (DXA) are all reliable methods to obtain accurate measures of total body fat. However, because of their costs in terms of time and money, these methods are not practical in large epidemiological studies and for routine clinical use. In these situations, body mass index (BMI) is often used and assumed to represent the degree of body fat. BMI, however, does not distinguish between fat mass and lean (non-fat) mass. For example, well-trained body builders have a very low percentage of body fat, but their BMI may be in the overweight range because of their large lean (muscle) mass. In addition, in the elderly and non-Caucasian populations, the relationship between BMI and body fatness may be different as compared with younger Caucasian populations.9–14 Another potential limitation of the BMI is that the distribution of fat over the body is not captured. Many studies have shown that an abdominal fat distribution, independent of overall obesity, is associated with metabolic disturbances and increased disease risk.15–23 An increased abdominal fat accumulation is largely caused by the accumulation of visceral (or intra-abdominal) fat (for distinction of these fat depots, see Figure 1). Owing to metabolic differences between different fat depots, they differ in their role of predicting metabolic disturbances and diseases. Table 1 summarizes the capability of the most commonly used methods to assess total adiposity and fat distribution. Abdominal obesity is usually assessed by the easily measured waist circumference, the waist-to-hip circumference ratio (WHR), or the less-commonly used sagittal abdominal diameter (SAD). By the use of sophisticated imaging techniques, such as magnetic resonance imaging (MRI) and computed tomography (CT), different fat depots can be distinguished at the waist level, and it has been shown that in particular the visceral fat depot is associated with metabolic disease risk.24–30 Because the SAD or waist circumference alone are more strongly correlated with visceral fat than the WHR,31–35 guidelines tend to focus on waist circumference to estimate disease risk as suggested by Lean et al.36 These widely used cut-points (i.e. 102 cm for men and 88 cm for women) were originally based on a replacement of the classification of BMI,36 but other cut-points have also been suggested on the basis of relationships with visceral fat area.37 These simple cut-points may be helpful for the classification of Caucasian adults, but they are probably less appropriate in other ethnic groups38 or older age groups.39 No large or consistent differences have been found between SAD and waist circumference in relation to visceral fat.31,33–35,40,41 Cross-section of the abdomen in which subcutaneous and visceral fat can be distinguished. Lighter-coloured areas are muscles, bones, and organs Capability of different body fat measurements to estimate total body fat and fat distribution CT, computed tomography; MRI, magnetic resonance imaging; DXA, dual-energy X-ray absorptiometry; BIA, bioelectrical impedance analysis; BMI, body mass index; WC, waist circumference; HC, hip circumference; WHR, waist-to-hip ratio; SAD, sagittal abdominal diameter. Capability of different body fat measurements to estimate total body fat and fat distribution CT, computed tomography; MRI, magnetic resonance imaging; DXA, dual-energy X-ray absorptiometry; BIA, bioelectrical impedance analysis; BMI, body mass index; WC, waist circumference; HC, hip circumference; WHR, waist-to-hip ratio; SAD, sagittal abdominal diameter. The waist circumference, however, is not always a stronger predictor of type 2 diabetes42,43 and other cardiovascular risk16,20,44–46 than the WHR. Because a higher WHR can be the result of a larger waist as well as a smaller hip, this suggests that a smaller hip (representing either lower muscle and/or fat mass at the hips) may be involved in determining the increased risk associated with an increased WHR. Hip circumference is more strongly associated with leg fat mass, but also with leg lean mass (particularly in men).47 Indeed, WHR has been related to both a larger trunk fat mass, and a smaller leg fat mass (both sexes) and leg lean mass (men).47 In another study, WHR has been related to both larger visceral fat and smaller leg muscle areas in men.48 Several studies have found that a smaller hip circumference (for a given waist circumference) is associated with an increased risk.43,49–56 Early studies on fat distribution often used skinfold measurements to characterize subcutaneous fat distribution.57 Subcutaneous fat patterns were described using Z-score patterns, ratios of skinfolds (e.g. triceps/subscapular skinfold ratio and ratio of the sum of trunk skinfolds to the sum of skinfolds on extremities) and principal components analyses. The most promising and simple indicator of central fat distribution using skinfolds seems to be the subscapular skinfold. A prospective study showed that the subscapular skinfold predicted coronary heart disease in men independently of BMI and other cardiovascular risk factors58 but studies that have compared the use of skinfold over and above the use of body circumferences or circumference ratios are rare. A few studies have compared the associations between skinfolds and circumferences in relation to components of the metabolic syndrome59,60 and these showed that correlations of waist circumference with risk factors were either similar or stronger compared with those of the subscapular skinfold. Misra et al. have proposed that the subscapular skinfold should be considered as an indicator of central fat distribution separately of the waist circumference.61 It is, however, difficult to say with certainty that the measurements of skinfolds improve substantially the prediction of disease independently of body circumferences. The composition of the body changes with age, and this may have serious implications for the interpretation of anthropometric data of older persons. First, older persons are generally shorter than younger persons owing to secular trends in height and owing to shrinkage of the spine because of vertebral bone loss, kyphosis, and scoliosis.62,63 Consequently, the BMI of older persons may be overestimated and it is, therefore, preferable to use a body height measured in early life to calculate the BMI of older persons. Second, with age the amount of lean body mass decreases, a process called sarcopenia.64 Sarcopenia occurs even in weight stable persons owing to a concomitant increase in the amount of body fat.65,66 The relationship between BMI and percentage of body fatness has been shown to be dependent on age.9 Data from the Rosetta Study show that older adults have, on average, more fat than younger adults at any given BMI.67 Prediction equations to estimate body composition from BMI will generally underestimate the amount of body fat in the elderly.68 Also, a similar skinfold thickness in young and older persons represents a higher percentage of body fatness in older persons.68 Age-specific prediction equations should be used to estimate body fatness from skinfold thickness.69 A recent study in older men and women (mean age 60.4 years) showed that the change in skinfold thickness over time did not predict change in body fat mass, suggesting its limited use in longitudinal studies.70 Finally, the distribution of body fat also changes with age because relatively more fat accumulates in the abdomen and less fat at the extremities.71,72 Waist circumference has been suggested as an indicator of overall body fatness in older persons.73 Changes in waist and hip circumference were better anthropometric predictors of change in body fat mass over a 10 year period compared with changes in skinfold thickness.70 However, in persons aged 70–79 years the explained variance in total body fat remained higher for BMI than for waist circumference.74 The waist circumference has been shown to be similarly correlated with the amount of visceral fat in young and older persons.75 In several studies, waist circumference was a better predictor of visceral fat in older persons compared with SAD or WHR.32,76 However, in very old men and women (>70 years) SAD was a better predictor of visceral fat compared with the waist circumference.74 In that study the BMI performed as good as the waist circumference in predicting visceral fat.74 An alternative, non-anthropometric method proposed to estimate the amount of visceral fat in older persons is subregional DXA.40 The prediction of visceral fat by DXA was better compared with the waist circumference but was similar to the SAD. For a given waist circumference, visceral fat has been shown to be higher in older persons compared with younger persons,76 suggesting that absolute levels of waist circumference should be interpreted differently in younger and older persons. Prediction equations for visceral fat generally include age.77 Whether the proposed waist circumference cut-points are useful for the prediction of cardiovascular disease risk factors in older persons is still unclear.78,79 Several studies have shown a race difference in the association between BMI and percentage of body fat. For a given BMI, Chinese, Malay, Indian, Taiwanese, and Indonesian men and women have a higher percentage of body fat compared with Caucasians.10–14 Differences in trunk-to-leg-length ratio, slenderness, and muscularity may contribute to these racial differences in the percentage of body fat–BMI relationship.80,81 No clear differences in the relation between percentage of body fat and BMI have been observed for African-Americans vs Caucasians.67 Based on these findings and the observed differences in the relation between BMI and disease risk, lower BMI cut-points have been suggested to define overweight and obesity for specific ethnic groups.10–12,82–84 However, the expert committee of the WHO has not redefined the cut-points for specific Asian populations, because available data do not necessarily indicate a clear BMI cut-off point for all Asian ethnic groups.85 Ethnic differences in the relation between waist circumference and visceral fat have frequently been reported. Asian ethnic groups generally have a smaller waist circumference compared with Caucasians, although this is not necessarily true for Asian emigrants who are generally affluent and have more generalized and abdominal obesity.38,86 Despite the smaller waist circumference, the visceral fat mass is higher for Asians compared with Caucasians and African-Americans.87,88 Moreover, for specific Asian groups disease risks may already be increased at a lower level of waist circumference, suggesting that a lowering of the waist circumference cut-points should be considered for these ethnic groups.38,82,83 For a similar waist circumference (or WHR) and BMI, African-Americans have a lower visceral fat mass compared with Caucasians.89–91 Although a larger amount of visceral fat is associated with a higher disease risk in both African-Americans and Caucasians, similar amounts of visceral fat are associated with different levels of metabolic risk factors in these groups.92 The cause of the more diabetogenic profile in African-Americans as compared with Caucasians is currently unknown.93 Secular changes in the prevalence of overweight and obesity as measured by BMI have been reported in many countries over the last decades.1 Several studies have reported on secular changes in waist circumference and WHR. In German adults and in British adolescents, stronger increases over time in the average waist circumference than in relative weight were observed.94,95 A more rapid secular increase in WHR than in BMI was shown in Swedish women.96 Lahti-Koski et al.97 observed a similar secular change but did not compare this to the time trend in BMI. A more recent study from Sweden observed a significant increase in BMI but not in WHR in the period from 1985 to 2002.98 A Dutch study showed similar secular changes in waist circumference as compared with BMI over a short period of time.99 Large increases in the waist circumference were shown in the US adult population from 1960 to 2000 in all categories of BMI.100 It has been suggested that the waist circumference is more sensitive to changes in energy balance (e.g. decreased physical activity) than the BMI.99 With decreasing physical activity muscle may be gradually replaced by fat mass and this may have less effect on overall body weight than on the waist circumference. The prevalence of abdominal obesity according to these cut-points has been reported for several countries. In the US the age-adjusted prevalence of abdominal adiposity has tripled in men from ∼13% in 1960–62 to 38% in 1999–2000. In women the prevalence increased from ∼19 to 60% over the same period.100 In The Netherlands the prevalence of abdominal obesity was ∼15% in men and 21% in women in 1993–97.99 These prevalences were similar to those of the US some 30 years earlier. The worldwide increase in the prevalence of (abdominal) obesity is alarming because of the associated disease risk, in particular type 2 diabetes and cardiovascular diseases. Several studies were conducted to compare the contribution of measures of overall obesity (BMI) and abdominal obesity (waist circumference, WHR, SAD) with disease risk. Overall, it can be concluded that persons with a BMI in the normal weight range can still be at increased risk of metabolic disturbances if the WHR or waist circumference is increased, and that the combination of a high BMI and a high WHR results in a particularly high risk of an unfavourable metabolic profile, type 2 diabetes, and cardiovascular diseases.15–23 In the elderly, few studies have directly compared BMI and waist circumference as predictors of metabolic abnormalities. The limited available data suggest that waist circumference and/or SAD are better indicators of cardiovascular risk factors in older men and women compared with BMI.21,101 However, in other studies waist circumference contributed little to the prediction of disease risk after BMI had been taken into account.102,103 Regarding the comparison of waist circumference and WHR as predictors of metabolic disturbances and the risk of cardiovascular diseases, results have been inconsistent. Some studies have found waist circumference a stronger correlate of metabolic risk factors and cardiovascular disease than the WHR,31,104,105 whereas others found no difference,106–108 or found that the WHR was superior.16,20,42–46 The last observation resulted in several studies that investigated the separate contributions of waist and hip circumferences to disease risk. These studies consistently show that a smaller hip circumference, for a given waist circumference, is related to an increased risk for metabolic disturbances,49–51,53 type 2 diabetes,43,50,52,54 and cardiovascular disease and mortality.52,55,56 These results suggest that the measurement of hip circumference can contribute to the prediction of cardiovascular disease risk. If waist circumference or WHR were compared with the SAD in the prediction of metabolic disturbances and disease risk, some studies found SAD a stronger correlate than waist circumference, WHR, and BMI.109–111 However, waist and SAD correlated similarly (but stronger than WHR) with metabolic variables in another study.31 In others, none of these three measures was superior in their association with metabolic disturbances or cardiovascular disease risk.106 Considering more sophisticated body composition measurements, numerous studies have shown a consistent and strong association of CT-measured visceral fat area in relation to metabolic or disease risk.24–30 Results from the literature are not consistent on whether abdominal subcutaneous fat area contributes to an unfavourable metabolic profile and cardiovascular disease, independently of visceral fat area.112 In the Health, Aging and body composition Study, a larger amount of abdominal subcutaneous fat was substantially and independently associated with higher glucose and lipid levels, although associations were stronger for visceral fat.30 This independent association of subcutaneous abdominal fat with metabolic risk factors agrees with findings in earlier smaller studies.113,114 Studies using DXA or CT to estimate fat and muscle content at the legs found that in particular more subcutaneous fat at the and to a muscle mass at the was associated with a more cardiovascular risk profile (for a given amount of abdominal the association of a smaller hip circumference (for a given waist circumference) with increased disease risk can be explained by larger fat mass. Results for measures of body fatness and risk of are more difficult to than results for disease risk. First, of can substantially for different and the of body fatness on these will be Second, the time for of body fatness on is to be than for on the of associations between and particularly to owing to that are related to increased can result in weight most studies of adiposity and have used BMI as a measure of body Because BMI can both fat and lean body mass, in lean body mass that may be associated with can the interpretation of results for BMI. Indeed, the association between BMI and that has been observed in some studies may the of lean mass and fat mass with risk of Overall, study findings have that obesity (i.e. BMI 30 increases the risk of However, results for lower levels of BMI in relation to have been less and of studies have contributed to in Studies of body fat distribution and have been limited to studies of anthropometric measures of body fat distribution. In several studies, larger waist larger larger larger and smaller hip were substantially associated with risk of after for BMI. Because BMI may also in lean body mass, that these independent associations are owing to for overall body However, measures of central fat distribution also remained associated with after for overall body fatness assessed by skinfold or bioelectrical These results are consistent with results for morbidity that suggest that given a degree of body it is preferable to have fat in the of the abdominal It should be that studies of body fat distribution and have been conducted in Large waist circumference was a stronger predictor of than BMI in but measure was associated with in owing to the limited of the Although the that obesity, in particular abdominal obesity, is an cause of metabolic disturbances is generally the are not It is widely that an role in the of type 2 levels in the are high in abdominal glucose and by muscle and other organs is several The will the glucose by but in many of the the In addition, will in organs or fat particularly in the muscle and in the but also in the of fat in may and which are the in the of type 2 fat is more sensitive to and less sensitive to as compared with subcutaneous fat. visceral fat is more to into the increased levels, which may to fat in and It has been however, that the contribution to levels of subcutaneous fat is probably larger because is more of However, the of from visceral fat into the that directly to the may cause increased and increased of visceral fat and and glucose in whereas of amounts of subcutaneous fat had little a study in suggests that of of visceral improve the metabolic epidemiological studies it is whether a larger abdominal subcutaneous fat mass also contributes to an increased disease risk, independently of visceral fat. In a study in of abdominal subcutaneous fat of subcutaneous abdominal fat to by to of and glucose levels after of subcutaneous are Abdominal subcutaneous fat can be into and subcutaneous It was that the amount of subcutaneous had a stronger association with than subcutaneous fat, which may be to differences in This difference between different of subcutaneous fat may have contributed to the results of total abdominal subcutaneous fat and changes in also described in a recent studies suggest that more subcutaneous fat in the for a given amount of abdominal fat, may be associated with a more cardiovascular risk It has been suggested that the fat depot a role by as a for in the are relatively to and have a high for The an role in the of from the and particularly in the fat depot has a relatively high activity and relatively low of and the is more to from the and is less to a result of in the fat in the and may be In with this potential of subcutaneous in glucose and subcutaneous caused metabolic disturbances in In the role of subcutaneous is by the observation that or is by fat and type 2 No studies have been conducted that compared the of in the with in the abdominal The increases in a amount of total body fat is These into in particular in the These have a better to which may result in fat and in and many and with activity and Some of these may be involved in the of in The of is probably also to the different associations of abdominal and subcutaneous fat depots with metabolic risk factors may be to differences in the of between the fat There are differences in of and between abdominal subcutaneous fat and visceral whereas the of differences in the of and is is differences in the of these between abdominal and subcutaneous fat. In addition, are probably many more differently by different fat depots, which influence metabolic more in this area is There are several factors that may influence body fat as well as the of metabolic disturbances and therefore, or the associations between these These factors include factors physical factors in and factors such as age and it has been shown that fat depots are also by the in to the earlier of a and that of the lipid accumulation et have proposed an and as a cause of body fat and the metabolic In it is by studies, studies of metabolic risk and studies of cardiovascular disease and that body fat distribution is for the risk of cardiovascular disease and mortality. trend studies have shown that is a consistent increase over time in the prevalence of obesity abdominal obesity, which is to contribute to a higher incidence of type 2 diabetes, cardiovascular disease, and mortality. Several methods are available to measure body and the largely on the For clinical it should be considered that usually no on body fatness is at all and the of overweight are often not by with their The of a measure of body fatness in clinical may already be an and the of the measurement is an For this BMI can be an measure of body fatness in However, waist circumference may be a simple that also on abdominal fat distribution and may be less by in lean mass. The WHR is more difficult to because it may an effect of larger waist as well as a smaller hip circumference. The SAD can be used of waist circumference but has not consistently been shown to be superior for the prediction of disease risk. For large epidemiological studies the BMI can most of the in body fatness on the age of the study However, many studies have shown that the of on body fat distribution (waist circumference, WHR, SAD, can For studies and studies with that may body fat distribution, accurate to assess fat depots MRI, is
DOI: 10.1093/ije/dyi253
Authors: MB Snijder, RM van Dam, M Visser, JC Seidell
Language: English
Subfield: Physiology (medical)
Field: Medicine
Domain: Health Sciences
Open Access status: bronze
OpenAlex ID: W2125930781