Background:

The health risks of obesity are disproportionately due to central abdominal adiposity; however, the extent to which age is associated with the body shape of obese adults is not known.

Objective:

Three-dimensional (3D) data on body shape from the UK National Sizing Survey were analyzed to investigate age-associated changes in body shape within the BMI bands <20, 20–24.99, 25–29.99 and 30 kg/m².

Methods:

Measurements of anthropometry (weight and height) and a 3D body scan were obtained in 4,344 men and 5,266 women recruited from eight British cities.

Results:

The body shape of men showed high consistency within BMI bands between early adulthood and old age. In contrast, the body shape of women altered within each BMI band with increasing age. In obese, overweight, and normal weight women, age was associated with decreased thigh girth, increased waist, and bust girth. Whereas young obese women maintained an hourglass shape, in old age the body shape of obese women converged on that of obese men.

Discussion:

The association of age with body shape is markedly different between the sexes, with the impact of obesity on shape strongly age-dependent in women but not in men. The age delay in the association between obesity and high waist girth in women may contribute to the sex-difference in life expectancy. The relationship between body shape change and cardiovascular risk merits longitudinal investigation within individuals.

Introduction

The categorization of obesity on the basis of BMI (calculated as weight/height² in kg/m²) (1) has facilitated epidemiological investigations of high levels of body weight, and enabled a simple and convenient assessment in clinical practice. Increasingly, however, some limitations of this index of obesity have become apparent.

It is now widely acknowledged that BMI cannot distinguish between lean and fat tissue (2,3). The index may categorize some relatively muscular individuals such as athletes as of excess body weight, and conversely fails to recognize individuals who have high body fat but low lean mass and hence are within the normal weight range. To address this limitation, many researchers now advocate measurement of waist girth to clarify the level of central adiposity (4,5). A number of studies have demonstrated that waist girth predicts the risk of cardiovascular disease of type 2 diabetes independently of BMI (6,7,8). In particular, a large multicountry study of 52,000 adults showed that a high waist girth within a low BMI category was associated with a greater risk of cardiovascular disease than a low waist girth within a high BMI category (9).

Such research illustrates both the value of incorporating shape indices in the categorization of risk, and implies that some fat depots are more harmful for health than others. Abdominal and visceral fat in particular have been implicated as major risk factors for the metabolic syndrome (10,11). Recent work has also suggested that adipose tissue in the thigh is protective against the metabolic syndrome (12), and that a high ratio of abdominal to thigh fat is most strongly predictive of cardiovascular disease (13,14) and type 2 diabetes (15).

Between-individual differences in fat distribution are therefore strongly implicated in differential disease risk. Three-dimensional (3D) body scanning, providing a detailed body shape map, has therefore been proposed as a sophisticated approach for categorizing obesity status of individuals, allowing the identification of those most at risk of cardiovascular disease (16,17). The same method is ideal for monitoring individuals over time, identifying regional changes in body shape. Such emphasis on the value of shape is not new—in the mid 20th century, Sheldon developed a conceptual approach to body type termed somatotyping, categorizing individuals on scales in terms of relative fatness, muscularity and linearity of physique, and testing for associations with psychological profile (18). Scientists have in general remained skeptical as to the value of this approach, particularly the psychological component, although Carter and Heath have since investigated somatotype in relation to other physical and genetic traits (19). 3D scanning might be considered to bridge this gap, providing more detailed information about physique whilst allowing extraction of outcomes already routinely investigated in clinical practice and research.

The UK National Sizing Survey obtained 3D body scans in more than 9,000 adults aged 18–80+ years. Initial analyses of these cross-sectional data showed a positive association between age and weight in both sexes, with this association being stronger in women (17). There were also associations between age and body girths, but these two effects were not considered independently. The aim of these analyses is to describe age-associated variability in the body shape characteristic of overweight and obese men and women.

Methods

The UK National Sizing Survey has been described in detail previously (17). In brief, almost 10,000 adults were recruited into a survey of body shape using 3D body scanners. Participants were recruited in eight UK cities, Birmingham, Cardiff, Edinburgh, Leeds, London, Manchester, Nottingham, and Southampton. The survey was a collaboration between a consortium from the UK retail industry, the UK Department of Trade and Industry, and eight regional universities under the overall coordination of University College London. Recruitment aimed to complete minimum numbers according to cells categorized by sex, age (16–25, 26–35, 36–45, 46–55, 56–65, 66–75, and 76+ years), and social class (A, B, C1, C2, D, and E) based on education and income criteria, ethnicity and geographic region. Recruitment was first open to all those eligible, and then targeted at incomplete cells. Individuals signed a consent form permitting their anonymized data to be used in scientific and commercial analyses. We subsequently obtained approval from the Research Ethics Committee of Great Ormond Street Hospital and the Institute of Child Health to analyze the data for medical research purposes.

Participants underwent a 3D body scan using TC2 instrumentation (Cary, North Carolina, USA; www.tc2.com), dressed in close-fitting underwear. The procedure involved the subject standing in a specified stance in a customized photo booth for a period of around 10 seconds whilst light stripes were projected on to the body skin surface and the distortions generated by body shape recorded by cameras. Computer algorithms automatically extracted a suite of around 200 measurements from each scan, using automatic landmark identification. For the current analyses, girths of the waist, hip, chest, bust (women only), mid-thigh, and mid–upper arm were considered. 3D scanning measurements are slightly but systematically different from those obtained using manual anthropometry, due firstly to minor inconsistency in the locations measured and secondly to the fact that manual measurements, but not 3D measurements, involve pressure on the skin. For example, mean difference in waist girth and hip girth was 1.3 and 5.7 cm, respectively. However, comparison of 3D and manual measurements shows extremely high ranking consistency between methods (r = 0.97 (17)), and epidemiological associations should not therefore differ between techniques. Manual measurements of weight and height were also obtained, as these are not provided by the 3D scan.

Statistics

Conventionally, body shape is assessed using individual girths, such as waist or mid–upper arm girth, or by simple ratios such as the waist–hip ratio. However, it is now widely appreciated that dividing one measurement by another fails to address allometric issues satisfactorily. Although the best-known example comprises BMI, calculated as weight/height² rather than weight/height, many other similar indices have recently been explored from a similar perspective. These are fat mass/weight (percentage fat) and the ratio of arm to trunk skinfold thickness ratios (20).

In these analyses, for each age group, each gender was divided up into the BMI categories <20, 20–24.99, 25–29.99, and >30 kg/m², termed "thin," "normal weight," "overweight," and "obese" for descriptive purposes. General linear models were then constructed, in which any given outcome (e.g, waist girth) was adjusted for another girth (e.g., hip girth). This approach generated values for the adjusted girth for each BMI category, which allowed graphic analyses of variability between BMI bands and the sexes. These analyses adjusted waist girth for hip, chest or thigh girth in both sexes, and bust girth for hip or thigh girth in women.

Results

Tables 1 and 2 present mean and standard deviation of BMI and height for each BMI group across the age categories in men and women, respectively. Across the age groups, the average BMI of overweight and obese individuals was very similar in each sex despite the increasing proportion of obese individuals with increasing age. The mean BMI of obese men varied only from 32.6 to 33.3 kg/m² in the seven male age categories, and from 33.5 to 34.9 kg/m² in the equivalent female agecategories. For thin, normal weight, and overweight bands of BMI, the range of mean BMI in all seven age categories was within 1 kg/m² for both sexes.

Table 1 - Height, BMI, and girths by age and BMI category in men.

Full table (99K)

Table 2 - Height, BMI, and girths by age and BMI category in women.

Full table (107K)

BMI was inversely associated with height in men in all age groups, except for 41–50 years when the P value was 0.07. In women, an inverse association between BMI and height was only significant in the age group 21–30 years. Tables 1 and 2 also present the raw data for girths in each BMI band in each age group for men and women, respectively. All associations between girths and BMI category were highly significant in all age groups in both sexes.

Figure 1a and b present plots by age of waist girth adjusted for hip girth in men and women, respectively. It can be seen that waist girth adjusted for hip girth increased across the BMI bands in men, but that age had relatively little effect on this trend. In contrast, both age and BMI band had a strong effect in women, with the oldest group of thin women having similar adjusted waist girth to the youngest group of obese women, these values in turn being similar to those of thin males of any age. In the oldest group of obese women, adjusted waist girth was similar to that of overweight men of all ages.

Figure 1.

Waist girth adjusted for hip girth by BMI category and age group in (a) men and (b) women, respectively.

Full figure and legend (35K)

Figure 2a and b present equivalent plots for waist girth adjusted for chest girth. Again, age had little influence on male waist girth adjusted for chest girth, but a strong effect in women. The youngest group of obese women had a similar adjusted waist girth to that of normal weight men, whereas the oldest group of obese women had a similar adjusted waist girth to that of obese men.

Figure 2.

Waist girth adjusted for chest girth by BMI category and age group in (a) men and (b) women, respectively.

Full figure and legend (32K)

Figure 3a and b show similar plots for waist girth adjusted for thigh girth. Here, thin women had similar waist girth adjusted for thigh girth to thin men of any age. However, the youngest group of obese women had a similar adjusted waist girth to that of normal weight men, whereas the oldest obese women had an adjusted waist girth midway between those of overweight and obese men.

Figure 3.

Waist girth adjusted for thigh girth by BMI category and age group in (a) men and (b) women, respectively.

Full figure and legend (31K)

Figure 4 shows the association between BMI category and bust girth adjusted for hip girth in each age group in women. The association of adjusted bust girth with BMI category was strongly influenced by both BMI category and age, with the oldest women in any given BMI category having the same adjusted bust girth as the youngest women in the next BMI category.

Figure 4.

Bust girth adjusted for hip girth by BMI category and age group in women.

Full figure and legend (19K)

Figure 5 shows the association between BMI category and waist girth adjusted for bust girth in each age group in women. The effect of age was very strong, such that the oldest group of thin women had a higher waist girth adjusted for bust girth than the youngest group of obese women. Thus, even relative to a region of body fat that itself increased with age, waist increased substantially more.

Figure 5.

Waist girth adjusted for bust girth by BMI category and age group in women.

Full figure and legend (18K)

Discussion

Increasingly, research has demonstrated that it is central body fat rather than weight relative to height per se that most strongly predicts ill-health and risk of mortality (6,7,8,9). Our analyses show that the body shape characteristic of overweight and obesity is similar across the adult lifespan in men, but markedly different in women. In normal weight, overweight, and obese women, increasing age is associated with a significant shift of body weight from the lower to the upper body, and in particular to the waist. Hence, young and older women within the same BMI had markedly different fat distributions, unlike in men, in whom shape variability could be attributed to BMI category and not age. Our study thus highlights the limitations of BMI as an index of cardiovascular risk across the adult lifespan, particularly in women.

A limitation of our study is that it relies on cross-sectional data, and hence cannot reveal actual changes in shape within the individuals during the life course. Young individuals today may not pursue the same changes in shape that occurred in older individuals due to their exposure to different environmental factors. Nevertheless, the UK National Sizing Survey was a large-scale survey allowing relatively large groups of individuals of each age group and sex to be assessed in relation to BMI category. The age-associated changes in body shape we have described in women are known to occur within individuals, in association with several factors including age itself, parity and the menopause (21,22,23). A further limitation is that we had no data on lifestyle factors such as diet or activity patterns, each of which is known to account for variability in shape in both sexes (25). However, our large sample with representative composition in each age and sex category reduces the likelihood of our findings being an artefact of differences in lifestyle as opposed to biological effects of age and gender.

Differences in body shape in early adulthood are well known (24,25), and can be attributed to sexual dimorphism related to reproductive function in combination with sexual selection (25,26). Our data show that during the life course sexual dimorphism largely disappears, so that by old age women within all BMI categories are characterized by a relatively android pattern of fat distribution. Although our previous analysis suggested this loss of dimorphism (17), the present analysis adds further information by demonstrating that the association between age and shape is observed across the entire spectrum of female body weight, and is not simply a by-product of the tendency for the average woman to increase in BMI over adulthood. The lack of such associations in men implies that hormonal correlates of female reproductive biology change the strategy of energy deposition with age, with concentrations of androgens and estrogen likely candidates (26).

Between-individual variability in the magnitude of shifts in adipose tissue distribution in women is likely to be associated with risk of the metabolic syndrome. An increased central fat distribution in both sexes is associated with increased risk of disease (6,7,8,9,13,14); however, the persistence of an hourglass body shape (low waist girth relative to hip and bust girths) across the entire range of BMI (including obesity) in young women may indicate that the toxic effects of abdominal fat are delayed in women, and begin to exert their effects at a later age compared with men. Coronary heart disease is now appreciated to have an incubation period (27), such that metabolic profile before the time of death is more predictive of that event that metabolic profile at the time of death. It is plausible therefore that the relative delay in women compared to men in the accumulation of central adiposity, regardless of BMI category, is a significant factor contributing to the sex difference in life expectancy and risk of cardiovascular disease at any given age. This hypothesis merits further attention, in particular in relation to our understanding of how the components of female reproductive history (pregnancy, lactation, and menopause) influence body fat distribution.

The negative correlation between BMI and height in men might appear to derive from an imperfect statistical adjustment of weight for height using this index. However, our previous analysis found an inverse association of height with waist girth, whereas equivalent associations with all other girths were positive (17). This suggests that there may be a biological association between poor growth and obesity, and this issue, particularly as it appeared sex-specific, merits further association.

Traditionally, anthropometry has been considered a relatively unsophisticated approach for evaluating body fatness. Certainly, BMI is a poor proxy for body composition, while skinfold thicknesses address only subcutaneous fat. Contemporary studies increasingly attempt to measure body fat by more direct approaches such as isotope dilution, dual energy X-ray absorptiometry, magnetic resonance imaging or computed tomography scanning. Of these, only the radiographic imaging techniques provide information on fat distribution, and only magnetic resonance imaging or computed tomography scanning do so with high regional accuracy. Despite these benefits, magnetic resonance imaging and computed tomography are not widely used due to their high cost, and in the case of computed tomography, high radiation exposure. Body shape offers detailed information about fat distribution, partly because much body fat is subcutaneous and therefore closely associated with superficial topography, and partly because dimensions such as waist girth and sagittal diameter show relatively good correlations with abdominal visceral and deep subcutaneous fat (28,29,30,31,32). Such information is likely to be of greater utility than a single value for whole body fat mass, which is often adjusted inappropriately for body size (20). Body surface dimensions should not therefore be regarded as a primitive approach for describing fat distribution, rather 3D body scanning has the potential to play a major role in both epidemiological studies of risk of the metabolic syndrome, and the monitoring of individual patients in response to treatment. We are currently developing a range of more detailed 3D shape outputs (33) that make possible a new approach to using body shape information as an index of health status.

Analysis of the UK National sizing survey demonstrates a marked difference between the sexes in the association between age and body shape. Younger women tend to have an hourglass fat distribution regardless of their BMI, whereas older women tend to have a more android fat distribution again regardless of their BMI. This association of age with shape suggests a delayed onset of the toxic effects of central fat in women and may contribute to the average greater lifespan of women in industrialized populations. 3D body scanning is a valuable technique for exploring associations between body shape and disease risk.

Disclosure

The authors declared no conflict of interest.

References

REFERENCES

1. Garrow JS, Webster J. Quetelet's index (W/H2) as a measure of fatness. Int J Obes 1985;9:147–153.  | PubMed | ISI | ChemPort |
2. Wells JCK. A Hattori chart analysis of body mass index in infants and children. Int J Obes 2000;24:325–329.  | Article | ISI | ChemPort |
3. Gurrici S, Hartriyanti Y, Hautvast JG, Deurenberg P. Differences in the relationship between body fat and body mass index between two different Indonesian ethnic groups: the effect of body build. Eur J Clin Nutr 1999;53:468–472.  | Article | PubMed | ISI | ChemPort |
4. McCarthy HD, Ashwell M. A study of central fatness using waist-to-height ratios in UK children and adolescents over two decades supports the simple message—'keep your waist circumference to less than half your height'. Int J Obes 2006;30:988–992.  | Article | ISI | ChemPort |
5. Ashwell M, Hsieh SD. Six reasons why the waist-to-height ratio is a rapid and effective global indicator for health risks of obesity and how its use could simplify the international public health message on obesity. Int J Food Sci Nutr 2005;56:303–337.  | Article | PubMed |
6. Rimm AA, Hartz AJ, Fischer ME. A weight shape index for assessing risk of disease in 44, 820 women. J Clin Epidemiol 1988;41:459–465.  | Article | PubMed | ChemPort |
7. Walton C, Lees B, Crook D et al. Body fat distribution, rather than overall adiposity, influences serum lipids and lipoproteins in healthy men independently of age. Am J Med 1995;99:459–464.  | Article | PubMed | ChemPort |
8. Meisinger C, Doring A, Thorand B, Heier M, Lowel H. Body fat distribution and risk of type 2 diabetes in the general population: are there differences between men and women? The MONICA/KORA Augsburg cohort study. Am J Clin Nutr 2006;84:483–489.  | PubMed | ChemPort |
9. Yusuf S, Hawken S, Ounpuu S et al. Obesity and the risk of myocardial infarction in 27,000 participants from 52 countries: a case-control study. Lancet 2005;366:1640–1649.  | Article | PubMed | ISI |
10. Kissebah AH, Vydelingum N, Murray R et al. Relation of body fat distribution to metabolic complications of obesity. J Clin Endocrinol Metab 1982;54:254–260.  | PubMed | ISI | ChemPort |
11. Bjorntorp P. Abdominal obesity and the development of noninsulin-dependent diabetes mellitus. Diabetes Metab Rev 1988;4:615–622.  | PubMed | ISI | ChemPort |
12. Kahn HS, Austin H, Williamson DF, Arensberg D. Simple anthropometric indices associated with ischemic heart disease. J Clin Epidemiol 1996;49:1017–1024.  | Article | PubMed | ISI | ChemPort |
13. Smith DA, Ness EM, Herbert R et al. Abdominal diameter index: a more powerful anthropometric measure for prevalent coronary heart disease risk in adult males. Diabetes Obes Metab 2005;7:370–380.  | Article | PubMed | ISI | ChemPort |
14. Snijder MB, Visser M, Dekker JM et al. Low subcutaneous thigh fat is a risk factor for unfavourable glucose and lipid levels, independently of high abdominal fat. The Health ABC Study. Diabetologia 2005;48:301–308.  | Article | PubMed | ISI | ChemPort |
15. Snijder MB, Dekker JM, Visser M et al. Associations of hip and thigh circumferences independent of waist circumference with the incidence of type 2 diabetes: the Hoorn Study. Am J Clin Nutr 2003;77:1192–1197.  | PubMed | ISI | ChemPort |
16. Wang J, Gallagher D, Thornton JC et al. Validation of a 3-dimensional photonic scanner for the measurement of body volumes, dimensions, and percentage body fat. Am J Clin Nutr 2006;83:809–816.  | PubMed | ISI | ChemPort |
17. Wells JCK, Treleaven P, Cole TJ. BMI compared with 3D body shape: the UK National Sizing Survey. Am J Clin Nutr 2006;85:419–425.
18. Sheldon WH. Atlas of Men: A Guide for Somatotyping the Adult Male at All Ages. Harper: New York, 1954.
19. Carter JE, Heath BH. Somatotyping: Development and Applications. Cambridge University Press: Cambridge, 1990.
20. Wells JCK, Victora CG. Indices of whole-body and central adiposity for evaluating the metabolic load of obesity. Int J Obes (Lond) 2005;29:483–489.  | Article | PubMed | ChemPort |
21. Blandeau TE, Hunter GR, Sirikul B. Intra-abdominal adipose tissue deposition and parity. Int J Obes 2006;30:1119–1124.  | Article |
22. Lassek WD, Gaulin SJC. Changes in body fat distribution in relation to parity in American women: a covert form of maternal depletion. Am J Phys Anthropol 2006;131:295–302.  | Article | PubMed |
23. Beaufrère B, Morio B. Fat and protein redistribution with aging: metabolic considerations. Eur J Clin Nutr 2000;54(Suppl 3):S48–S53.
24. Paeratakul S, Adair LS, Zhai F, Ge K, Popkin BM. Sex difference in measures of body fatness and the possible difference in the effect of dietary fat on body fatness in men and women. Eur J Clin Nutr 1999;53:865–871.  | Article | PubMed | ChemPort |
25. Wells JCK. The evolution of human fatness and susceptibility to obesity: an ethological approach. Biol Rev 2006;81:183–205.  | Article | PubMed |
26. Wells JCK. Sexual dimorphism of body composition. Best Pract Res Clin Endocrinol Metab 2007;21:415–430.  | Article | PubMed |
27. Rose G. Incubation period of coronary heart disease. BMJ 1982;284:1600–1601.  | PubMed | ChemPort |
28. Chan DC, Watts GF, Barrett, PHR, Burke V. Waist circumference, waist-to-hip ratio and body mass index as predictors of adipose tissue compartments in men. Q J Med 2003;96:441–447.  | ISI | ChemPort |
29. Ashwell M, Cole TJ, Dixon AK. Ratio of waist circumference to height is strong predictor of intra-abdominal fat. BMJ 1996;313:559–560.  | PubMed | ChemPort |
30. Brambilla P, Bedogni G, Moreno LA et al. Crossvalidation of anthropometry against magnetic resonance imaging for the assessment of visceral and subcutaneous adipose tissue in children. Int J Obes 2006;30:23–30.  | Article | ISI | ChemPort |
31. Zamboni M, Turcato E, Armellini F et al. Sagittal abdominal diameter as a practical predictor of visceral fat. Int J Obes 1998;22:655–660.  | Article | ISI | ChemPort |
32. van der Kooy K, Leenen R, Seidell JC, Deurenberg P, Visser M. Abdominal diameters as indicators of visceral fat: comparison between magnetic resonance imaging and anthropometry. Br J Nutr 1993;70:47–58.  | Article | PubMed | ChemPort |
33. Wells JCK, Ruto A, Treleaven P. Whole-body three-dimensional photonic scanning: a new technique for obesity research and clinical practice. Int J Obes (Lond) 2007;e-pub ahead of print 9 Oct 2007.

Acknowledgments

The UK National Sizing Survey was funded by a consortium of UK industry and the UK Department of Trade and Industry.