Central overweight and obesity in British youth aged 11–16 years: cross sectional surveys of waist circumference

Introduction

Recent studies have shown an increase in the prevalence of overweight and obesity in British children over the past 10–15 years. 1 2 These findings have been based on body mass index and recently standardised cut-off points. 3 4 But body mass index gives no indication of the distribution of body fat, and in children, as in adults, centralised or upper body fat carries an increased risk for metabolic complications. For example, visceral adipose tissue is significantly related to concentrations of plasma low density lipoprotein cholesterol and triglycerides in 11–15 year olds as well as to concentrations of basal insulin and high density lipoprotein cholesterol (inversely) in females aged 10–16 years. 5 6 Furthermore, abnormalities in lipid concentrations and insulin concentration correlate with waist circumference in young people. 7 8 Waist circumference is a highly sensitive and specific measure of upper body fat in young people and so should be valuable for identifying overweight and obese children at risk of developing metabolic complications. 9 10 To this end we have recently had published the first waist circumference centile curves for British children and adolescents.11

An examination of secular trends in waist circumference in British children would add valuable comparative data to the evidence of trends in body mass index. We compared data on waist circumference, together with body mass index, in the same young people, collected in surveys over the past 10–20 years in Great Britain.

Methods

Results

The mean waist circumference and body mass index for females and males by age were appreciably greater in the national diet and nutrition survey than in the British Standards Institute survey (P<0.001 for all groups except body mass index in males aged 13, P<0.05), the differences being similar at all ages (figs 1 and 2). The mean increase in body mass index was 1.5 (SE 0.16) for males and 1.6 (SE 0.16) for females. The mean increase in waist circumference was 6.9 (SE 0.46) cm for males and 6.2 (SE 0.34) cm for females (all P<0.0001). The standard deviations were also larger in the national diet and nutrition survey than in the British Standards Institute survey.

Fig 1

Waist circumference in British youth aged 11–16 years. *P<0.0001

• Download figure
• Open in new tab
• Download powerpoint

Fig 2

Body mass index in British youth aged 11–16 years. *P<0.05, †P<0.001

• Download figure
• Open in new tab
• Download powerpoint

Standard deviation scores for waist circumference increased much more than for body mass index. The mean increase in waist circumference in males was 0.84 units versus 0.47 units for body mass index, whereas in females the contrast was even greater (1.02 units versus 0.53; table 1). Waist circumference increased by 0.18 (SE 0.08) units more in the females than in the males, and the variability in the females in 1997 was also significantly greater than that in the males. The changes in the females correspond to a shift from the 50th to the 85th centile for waist circumference and the 44th to the 70th centile for body mass index.

The standard deviation scores for waist circumference were strongly correlated with body mass index (r=0.82 in males, 0.79 in females) and less with height (r=0.48 and 0.30). Adjusting for body mass index and height changes between surveys reduced the increase in mean waist circumference from 0.84 to 0.45 (SE 0.029) units in males and from 1.02 to 0.57 (SE 0.033) units in females. So less than half the increase in waist circumference was explained by increases in body mass index or height.

Table 2 shows the prevalence by sex of overweight and obesity based on body mass index and waist circumference in the two surveys. In all cases the prevalence increased over time—for example, for overweight from 6-9% in 1977 and 1987 to 17-38% in 1997 (P<0.0001). The increases were appreciably larger for waist circumference (20-29% overweight and 11-14% obese) than for body mass index (11-13% overweight and 7% obese) (P<0.001). Based on waist circumference, more than one third of females in 1997 were overweight and over one sixth were obese.

Discussion

We examined two indices of fatness, body mass index and waist circumference, in British youth aged 11–16 years surveyed 10 years apart (1987 and 1997 for females) and 20 years apart (1977 and 1997 for males). We found large increases in both indices, substantially more for waist circumference than for body mass index, reflecting the widely documented trend to increased child obesity. This suggests a steeper rise in abdominal obesity than whole body obesity based on weight and height.

The definition of child overweight and obesity is arbitrary.4 The cut-off points used here, the 91st (overweight) and 98th (obese) centiles, were chosen because they are standard on the British charts; and because the 91st and 98th centiles for body mass index are similar to the recently described cut-off points for overweight and obesity determined by the International Obesity Task Force.4

The increased waist circumference is unlikely to be due to visceral adipose tissue alone; it probably reflects both visceral and subcutaneous fat and hence total fatness. In contrast the body mass index measures the sum of fat mass and fat free mass, and it is impossible to know the relative contributions of each.15 This suggests that body composition has changed over time, the increase in fat mass being obscured by a reduction in muscle mass. This has been shown by comparing contemporary children with Fomon's reference child.16 Monitoring trends in body mass index alone fails to identify this shift from muscle to fat.

In obese adolescents the accumulation of excess fat is known to occur predominantly in the upper body rather than in the peripheral region. 6 17 A similar trend to central fatness based on measurements of skinfold thickness has recently been observed in Spanish youth, showing that the phenomenon is not unique to Britain.18

Methodological differences may have accounted in part for these observations. However, in both surveys the site of the waist measurement was defined identically. Furthermore, with a low reported technical error of measurement,19 low intraobserver and interobserver error (unpublished data), and adjustment for clothing, it is likely that the increases in waist circumference are genuine.

It is not yet possible to guess the impact of these increases in waist circumference on current and future morbidity, but they should be a cause for concern. One US study found that young people above the 90th centile for waist circumference had higher concentrations of low density lipoprotein cholesterol, triglycerides, and insulin and lower concentrations of high density lipoprotein cholesterol than young people below the 10th centile.8 This may well be mirrored in British youth. It is not known how early in life the increases in waist circumference over the reference values from 1977 and 1987 are detectable in a contemporary population, although we have observed similar increases in children as young as 3 years (unpublished observations).

Another concern is the difference between the sexes. Males were followed over 20 years compared with 10 years for females, so the males ought to have shown a larger secular trend. Yet the increase was appreciably larger in the females than in the males, for reasons that are unclear. Energy intake has decreased to a similar extent in the sexes.13 It may be that levels of physical activity have decreased faster in females than in males. If so then central fatness may be related more to physical activity than to energy intake.

In conclusion, we found that in British youth aged 11–16 years, waist circumference, representing central fatness, has increased much faster than body mass index over 10–20 years. Further work is necessary to examine the link between waist circumference, body fatness, and morbidity in young people. Waist circumference should be routinely measured in schoolchildren. Urgent action is needed to reverse the trends and reduce the likely medical and psychological costs of the increase in obesity in young people.