Introduction. KR Fox 1 *, DM Peters 2, P Sharpe 3 and M Bell 3

Transcription

1 (2000) 24, 1653±1659 ß 2000 Macmillan Publishers Ltd All rights reserved 0307±0565/00 $ Assessment of abdominal fat development in young adolescents using magnetic resonance imaging KR Fox 1 *, DM Peters 2, P Sharpe 3 and M Bell 3 1 Department of Exercise and Health Sciences, University of Bristol, Bristol, UK; 2 Physical Education Unit, Devon and Cornwall Constabulary, Exeter, UK; and 3 Somerset Magnetic Resonance Imaging Centre, Marsh Lane, Bridgwater, Somerset, UK OBJECTIVE: To determine the patterns of change and the best anthropometric indicators of intra-abdominal fat deposition in young adolescents from ages 11 ± 13 y. SUBJECTS: Subjects were 25 boys (mean age y) and 17 girls (mean age of y) who had taken part in a similar study 2 y earlier at ages y and y, respectively. METHODS: Intra-abdominal (IA) and subcutaneous adipose (SA) tissue areas and IA=SA ratio were determined through four tranverse magnetic resonance imaging scans on two occasions. Differences were investigated using t-tests and ANOVA. Skinfolds, girths and circumferences, body mass index and hydrostatic weighing were also recorded. Pearson correlation coef cients and regression equations were calculated to determine the best anthropometric indicators of intra-abdominal fat deposition. RESULTS: Intra-abdominal fat and subcutaneous fat areas had signi cantly increased in boys and girls by the second measure. Boys had deposited greater amounts of fat in intra-abdominal depots so that their intra-abdominal=subcutaneous ratio had increased signi cantly from 0.31 to This had reduced in girls from 0.39 to However, patterns of change were variable within sexes. Truncal skinfold sites (r ˆ 0.54 ± 0.70) emerged as the best eld indicators of intra-abdominal fat deposition. CONCLUSIONS: Patterns of intra-abdominal and subcutaneous fat distribution are identi able in pubescent children using magnetic resonance imaging. An acceptable indication is provided by truncal skinfolds. (2000) 24, 1653±1659 Keywords: body fat distribution; intra-abdominal adiposity; children; adolescents; magnetic resonance imaging; anthropometry; skinfolds; body mass index Introduction The health risks associated with excessive intraabdominal or visceral adipose tissue deposition have become increasingly well established. 1,2 This form of fat patterning is now seen as one of the major contributors to coronary heart disease and diabetes. The associated risks are independent of other factors such as overall levels of fatness, family history of disease and sedentary lifestyles. Furthermore, the risks emerge early in the lifespan and have been identi ed in obese adolescents. 3,4 Until the last 10 y, there was little direct measurement of intra-abdominal fat deposition either in adults or children. Proxy measures of central fat patterning using surface indices and ratios such as girths, circumferences or skinfold thicknesses were applied. *Correspondence: KR Fox, Department of Exercise and Health Sciences, University of Bristol, Bristol, BS8 1TN, UK. K.R.Fox@bristol.ac.uk Received 10 January 2000; revised 10 July 2000; accepted 2 August 2000 Trunk to limb skinfold ratios and the waist=hip circumference ratio have been the most favoured. Computed tomography (CT) has provided a more direct method of intra-abdominal fat deposition assessment with both adult and child populations. However, magnetic resonance imaging (MRI), which provides a radiation-free alternative, has been increasingly accepted as a safer alternative with children, particularly where multiple scans are required. It has provided reliable estimates of intra-abdominal fat and subcutaneous fat areas in transverse scans and recently, with a sample of men, multiple scans have shown promise in the estimation of fat volumes. 5 Progress has been made using CT or MRI with the estimation of intra-abdominal fat in pre-pubertal children, 6±8 early pubertal children, 9,10 and adolescents. 3,11 These studies, although few in number, have provided the rst assessments of the degree and nature of intra-abdominal deposition in children. They have also allowed the identi cation of indirect measures using skinfolds, circumferences and girths that are valuable for the estimation of abdominal adiposity with larger sample sizes in the eld. We are not aware of studies that have conducted repeated MRI measures with children across the

2 1654 pubertal years. The early origins of sex dimorphism found in adults with regard to intra-abdominal fat patterning have therefore not been documented. As a result, little is known about the nature of change of intra-abdominal fat patterning as childrens' bodies begin to develop adult characteristics, or how abdominal deposition tracks over this critical period. Furthermore, there is little information regarding the best indicators of abdominal deposition in early adolescence. The current study is a repetition of our previous work using MRI with a group of fty 11-y-old boys and girls 9,10 but over 2 y later when they had become 13-y-olds. The speci c purposes of the study were threefold. First, an assessment was made of the extent of total fat change during the period between the two studies. Second, an attempt was made to document the extent and nature of change in intra-abdominal adipose tissue in relation to other fat deposition. Third, the most useful anthropometric indicators for the prediction of intra-abdominal adipose tissue in 13-y-old adolescents were identi ed and compared to similar ndings in the previous study as 11-y-olds. Methods Subjects Fifty children (25 boys and 25 girls) who had been involved in a similar study at a mean age of 11 y 5 months 9,10 were invited to take part in this follow-up study at age 13. All the boys took part again. Eight of the 25 girls declined, representing a 32% attrition rate. The sample used in the current research therefore was 25 boys (mean age y) and 17 girls (mean age y). The mean age change was 2 y 2 months. Five of the eight missing girls had been in the lower two BMI quintiles in the rst study, suggesting that the new sample might be slightly fatter. However, a post hoc comparison using study 1 data revealed the same mean sum of four skinfolds for the 17 remaining as the original 25 girls. Procedures Ethical clearance was granted by the local Health Authority and informed written consent was obtained from the children and their parents=guardians for all the laboratory tests and the MRI procedures. Children were assessed in groups of six to eight over a 2 day period. They were asked to refrain from eating and drinking (except water) for 12 h prior to their involvement in any of the tests. Magnetic resonance imaging MRI scans were conducted with a 0.5 T superconducting MR system (Gyroscan T5=XPA, Philips Medical Systems, Philips, The Netherlands). A transmit= receive body coil was utilized with T 1 weighted spin echo sequences obtaining a 150 ms repetition time, a 20 ms echo time and a 10 mm slice width. The choice of the T 1 weighted spin echo sequences, repetition times and echo times were also within the ranges speci ed by Dooms et al 12 as being the most effective at the time for the imaging of adipose tissue. This was an upgraded machine from the rst study but parameters such as slice thicknesses and signal strength were consistent. Despite the similarities in scanning procedures, it must be noted that there has been no published research to indicate the consistency of different machines or different versions of software using MRI. In terms of reliability, however, all of the scans and data processing in both studies were performed by the same team of operators. Four transverse T 1 scans were performed on each child at the same anthropometric locations as the rst study. These were: mid-chest=mid-upper arm (level T8), waist (level L4 ± L5), hips (level of the greater trochanters) and mid thigh (mid femoral shaft) and were chosen as they incorporated some of the commonly used skinfold measurement sites. However, reporting here focuses on the waist scan at L4 ± L5 as it provides the main display of intra-abdominal fat. Fat areas were outlined and estimated using the free-standing Gyroview console (Philips, The Netherlands). The region of interest (ROI) was selected and the computer automatically shaded the rest of the pixels on the image which corresponded to their intensity. Manual adjustments of the boundaries of the ROI were made according to recommendations by Ross et al 13,14 to produce the best t area. Attempts were unsuccessfully made to outline and partition omental and mesenteric intra-abdominal deposition. Areas in cm 2 were calculated using the statistics option on the Gyroview. For each scan, total area (TA) and subcutaneous adipose tissue area (SA) were calculated. Additionally at the waist scan, intraabdominal adipose tissue area (IA) was calculated. Non-adipose tissue area (NA) for each scan site was calculated by subtracting the adipose tissue area of the scan from the total area of the scan. Grand total scan area, total adipose tissue area and total non-adipose tissue area were calculated by summation of the four scans areas. Finally, using the waist scan, IA=SA and IA=TA ratios were calculated for each child to provide an indication of the proportion deposited as intraabdominal fat. Skinfolds, anthropometry and maturation A comprehensive anthropometric survey was performed on each child. Height was recorded to the nearest millimetre using a xed base stadiometer. Weight was recorded using sliding beam scales to the nearest 10 g. Skinfolds were measured in triplicate with mean scores used, at locations suggested by Lohman et al. 15 These included biceps, triceps, subscapular, vertical and horizontal abdominal, suprailiac, mid thigh, medial calf and chest (boys

3 only). Circumferences were measured using a 0.7 mm wide metal tape at the waist, hip, mid thigh and chest (boys only) sites. Body breadths were measured at the biacromial and bitrochanteric sites using a sliding beam anthropometer. The same skinfold, girth and circumference combinations, including sum of skinfolds, trunk to limb ratios, waist ± hip circumference ratio (WHR) were used as in the previous study 9,10 to assess fat distribution. These combinations were selected from suggestions from previous research assessing childrens' fat distribution. 11,15,16 Finally, subjects were assessed for Tanner's stages of sexual maturation 17 by a trained nurse. All measures were taken by the same researcher using the same stadiometer, anthropometer, scales and skinfold calipers as the rst study. These are regularly checked through routine laboratory calibration procedures. Hydrostatic weighing Subjects were hydrostatically weighed within 24 h of the MRI scans, and had been asked to refrain from eating and drinking (except water) for 12 h prior to their appointment. Residual volume was measured using the Volugraph (Minjhardt, The Netherlands) through the helium dilution technique with the subject in the seated position. Each subject performed 20 fully submerged underwater weighing trials. The weight used as the child's underwater weight was selected as the heaviest weight from three consecutive readings within 0.1 kg of each other. Body density was converted to percentage fat using the age and sex-speci c Siri formulae adapted by Lohman et al. 15 One girl and one boy were unable to satisfactorily perform the underwater weighing procedure. Analyses The Statistical Package for the Social Sciences for Windows Release was used with statistical signi cance set at P < 0.05 for all analyses unless otherwise stated. Data are presented as mean standard deviation. Paired t-tests and two (age group) by two (sex) ANOVAs for were used to calculate means differences in key parameters between studies 1 and 2. Individual changes in body mass index, sum of skinfolds, and fat areas at the waist scan were also calculated along with percentage change to investigate differences in patterns of change among subjects of each sex. Finally, Pearson product moment correlation coef cients were calculated between all anthropometric variables and intra-abdominal fat deposition. Those variables showing strongest associations were used as independent variables in stepwise multiple regression analyses to determine the best predictors and standard errors of estimate (SEE) of intra-abdominal fat. These results were compared with study 1 ndings. Results Body fatness changes Group changes for the main growth and fatness parameters are summarised in Table 1. Boys and girls had reached 3.8 and 3.6 on Tanner's maturity indices and so had undergone considerable physical change. Body mass index increased in line with normal increases for growth. At the times of studies 1 and 2 the mean for boys was at the 60th percentile for their age using 1990 UK BMI references. 19 The girls BMI mean had increased slightly from the 55th to the 60th percentile between studies. Mean percentage body fat assessed by underwater weighing increased by 2.9% in boys and 6.3% in girls although the latter gure was based on only 14 girls with both sets of complete data. Increased fatness was re ected in an increase in total adiposity from all four MRI scans in both boys and girls. Total scan area increased by 13% for boys but adipose tissue area increased by 17%. Similarly, total scan area for girls increased by 18% for girls whereas adipose tissue area increased by 40%. Comparable increases in nonadipose tissue were 12% for boys and 7% for girls. This provides evidence of increasing sexual dimorphism by age 13 with girls depositing greater amounts and proportions of their weight as fat when compared to boys. Surprisingly, as most adipose tissue is still located subcutaneously at this age, skinfold measures did not re ect these increases with sum of four skinfolds showing no signi cant change in boys or girls. The only individual skinfold sites to show signi cant change were abdominal and thigh skinfolds which indicated small reductions in boys. This suggests that 1655 Table 1 Mean changes in size, maturity and fat-related variables for boys and girls Boys (n ˆ 25) Girls (n ˆ 17) Study 1 Study 2 Study 1 Study 2 Height (m) 1.47 (0.06) 1.62 (0.09)** 1.47 (0.08) 1.58 (0.08)** Weight (kg) 38.4 (6.0) 50.9 (9.4)** 39.9 (6.8) 50.8 (8.4)** BMI (kg=m 2 ) 17.8 (2.1) 19.4 (2.6)** 18.4 (2.9) 20.2 (3.0)* Tanner stage 1.0 (0.2) 3.8 (0.8)** 1.3 (0.60) 3.6 (0.50)** UWW percentage fat 13.6 (7.6) 16.5 (7.4)* 19.4 (8.5) 25.7 (6.6)** Sum of four skinfolds (mm) 40.0 (17.1) 36.8 (17.6) 48.5 (25.4) 48.0 (21.6) Total fat (from four MRI scans, cm 2 ) 357 (152) 419 (190)** 420 (192) 590 (213)** *P < 0.05; **P < 0.01.

4 1656 skinfolds are not suf ciently sensitive to re ect this degree of change in adiposity. Abdominal fat changes By study 2 there was also evidence of emerging sexual dimorphism with regard to abdominal fat deposition. Table 2 shows the areas of adipose tissue at the waist scan for boys and girls and also the percentage change in fat distribution. For boys, there was a mean percentage increase in waist scan area of 8%. This was accompanied by an increase of 69% in intra-abdominal fat area but only 19% increase in subcutaneous fat area. There was no mean increase in non-adipose tissue area so that the 8% increase in cross-sectional area and 10% increase in waist circumference was due to additional intra-abdominal fat deposition (12.3 cm 2 ) and subcutaneous fat deposition (14.9 cm 2 ). This was accompanied by a signi cant increase in IA=SA ratio from 0.31 to 0.39 and an increase in the percentage of total fat that is deposited as intra-abdominal fat. The pattern for females was different. There was an increase of 17% in waist cross-sectional area. Intraabdominal fat area signi cantly increased by 48%. However, there was a 78% increase in subcutaneous Table 2 Changes in fat deposition using the MRI waist scan at L4 ± L5 Boys (n ˆ 25) Girls (n ˆ 17) Study 1 Study 2 change Study 1 Study 2 change Total scan area 308 (66) 333 (70)** (72) 352 (79)* 16.9 Intra-abdominal fat area 17.8 (10.1) 30.1 (11.0)** (10.1) 38.3 (9.9)** 48.4 Subcutanous fat area 78.1 (48.5) 93.0 (55.1)** (42.4) (72.8)* 78.1 IA=SA ratio 0.31 (0.16) 0.39 (0.09)** (0.16) 0.35 (0.17)** 10.3 Intra-abdominal fat as percentage 5.6 (3.2) 7.2 (2.7)* (2.3) 6.4 (2.0) 3.1 of total MRI fat from four scans Waist circumference (mm) 636 (52) 698 (63)** (65) 671 (65)** 7.5 Areas in cm 2.*P < 0.05; **P < Table 3 BMI, sum of four skinfolds, and intra-abdominal fat change in boys BMI (study 1 rankings rst) (kg=m 2 ) Sum of four skinfolds (mm) L4 ± L5 Intra-abdominal L4 ± L5 subcutaneous IA=SA ratio ID Study 1 Study 2 change Study 1 Study 2 change Study 1 Study 2 Study 1 Study 2 Study 1 Study 2 B B B B B Mean Change 76.% 45.% 2.% B B B B B Mean Change 37.% 48.% 29.% B B B B B Mean Change 50.% 7.% 50.% B B B B B Mean Change 113.% 28.% 35.% B B B B B Mean Change 72.% 12.% 80.%

5 adiposity with the result that the IA=SA ratio signi cantly reduced by 10% from 0.39 to Although there was a greater absolute area of intra-abdominal fat in girls than boys, it also represented a signi cantly smaller percentage of total fat as measured by the four MRI scans (7.2% vs 6.4%, respectively). Although these changes are small, they are consistent with a shift to greater intra-abdominal or android deposition in boys with girls depositing more fat subcutaneously as in gynoid-type distribution. It is important to note that the total amount of intraabdominal fat remains small in both boys and girls, representing only 7.7% of total MRI adiposity and 9% of waist cross-sectional area. Within sex there is considerable variation in the pattern of body fat deposition and change, suggesting that genotyping is strongly in uential. Some insight into variations in fat pattern development can be gained by viewing key descriptive data on individuals. These are presented for boys in Table 3 and girls in Table 4. The data are listed by the BMI quintile groups at the time of their selection for study 1 so that directions of change for individuals can be more easily examined. Subject B3, for example, was originally in the leanest BMI quintile but has experienced a 13% increase in BMI and 52% increase in sum of four skinfolds. He initially had the least intraabdominal fat in the sample but has since moved into the upper ranges with an IA=SA ratio of Subject B16 has increased BMI by 12% and skinfolds by 40% but has maintained a low intra-abdominal fat deposition and lowered his IA=SA ratio. Subject B25 had the highest BMI in both studies but experienced a decrease in skinfold fat of 18% and also a small reduction in intra-abdominal fat and IA=SA ratio. In viewing girls' individual data, subject G10 increased BMI by 18%, skinfolds by 39%, but this was only re ected in increases in subcutaneous deposition at the waist resulting in a reduced IA=SA ratio from 0.35 to The fattest girl in study 1, as indicated by the highest BMI and sum of skinfolds, lost a lot of weight by study 2. Her BMI reduced from 25.3 to 18.0 kg=m 2 and she experienced a 67% decrease in sum of four skinfolds. However, the IA=SA ratio increased as the larger amount of fat was lost as subcutaneous tissue. Clearly, although there is evidence of sexual dimorphism, there is also considerable individual expression within the sexes Table 4 BMI, sum of four skinfolds and intra-abdominal fat change in girls BMI (study 1 rankings rst) (kg=m 2 ) Sum of four skinfolds (mm) L4 ± L5 Intra-abdominal L4 ± L5 subcutaneous IA=SA ratio ID Study 1 Study 2 change Study 1 Study 2 change Study 1 Study 2 Study 1 Study 2 Study 1 Study 2 G G G Ð 27.3 Ð Ð 16 Ð 40 Ð 0.40 Ð G Ð 30.8 Ð Ð 25 Ð 51 Ð 0.49 Ð G Mean Change 61.% 47.% 13.% G Ð 37.6 Ð Ð 22 Ð 62 Ð 0.36 Ð G Ð 40.0 Ð Ð 18 Ð 61 Ð 0.30 Ð G G Ð 42.8 Ð Ð 20 Ð 136 Ð 0.15 Ð G Mean Change 52.% 104.% 25.% G G G G G Ð 70.8 Ð Ð 23 Ð 152 Ð 0.15 Ð Mean Change 77.% 109.% 15.% G G Ð 57.4 Ð Ð 26 Ð 123 Ð 0.21 Ð G G G Ð 87.3 Ð Ð 32 Ð 122 Ð 0.26 Ð Mean Change 135.% 60.% 37.% G G G G G Mean Change 23.% 131.% 24.%

6 1658 Indicators of intra-abdominal fat Pearson product moment correlation coef cient between the main anthropometric variables and intra-abdominal fat area are presented in Table 5. Our previous study indicated that the best predictors of intra-abdominal fat deposition in 11-y-old girls were skinfold measures (sum of four skinfolds r ˆ 0.82, subscapular ˆ 0.83) suggesting that internal fat mirrored general levels of subcutaneous fat. A different pattern emerged for 13-y-olds. Coef cients were much lower, ranging from r ˆ 0.37 to 0.64, with the supscapular skinfold providing the strongest relationship. Besides skinfold measures, the only other signi cant association was body mass index (r ˆ 0.52) which may have been in ated because this was originally used in the strati cation of the sample. For boys, the reverse pattern emerged. With 11-yold boys, skinfold thicknesses, particularly limb skinfolds, were weakly to moderately associated with abdominal fat deposition (sum of four skinfolds r ˆ 0.43, subscapular ˆ 0.59). However, by 13 y of age, skinfolds were much stronger predictors (sum of four skinfolds r ˆ 0.72, abdominal horizontal r ˆ 0.70). Trunk to limb skinfold ratios, often used to estimate central fat deposition in children, were not useful indicators of intra-abdominal fat area in boys or girls. The subscapular=triceps ratio which had explained 41% of intra-abdominal deposition in boys in study 1 was no longer signi cantly correlated. Waist, hip and thigh circumferences were moderately related to intra-abdominal fat in boys. However, waist ± hip ratio, which has been widely used in Table 5 Correlations between waist intra-abdominal adipose tissue area and anthropometric variables with comparable study 1 gures (in parentheses) Parameter Boys (n ˆ 25) Girls (n ˆ 17) Height 0.10 (0.01) 0.16 ( 0.19) Weight 0.35 (0.38) 0.37 (0.48) BMI 0.58** (0.51**) 0.52* (0.67**) Maturity index 0.26 (0.04) 0.05 (0.11) Sum of four skinfolds 0.72** (0.43*) 0.53* (0.82**) Trunk skinfolds Chest 0.65** (0.53**) Ð Ð Subscapular 0.67** (0.59**) 0.64** (0.83**) Suprailiac 0.62** (0.40*) 0.60* (0.80**) Abdominal horizontal 0.70** (0.39) 0.54* (0.81**) Abdominal vertical 0.68** (0.43*) 0.56* (0.82**) Limb skinfolds Biceps 0.64** (0.37) 0.47 (0.81**) Triceps 0.64** (0.22) 0.55* (0.76**) Thigh 0.55** (0.38) 0.49* (0.69**) Calf 0.43* (0.28) 0.37 (0.74**) Skinfold formulas Subscapular=triceps 0.26 (0.64**) 0.34 (0.52**) Triceps biceps 0.65** (0.28) 0.53* (0.76**) Subscap suprailiac 0.65** (0.49*) 0.62** (0.79**) Circumferences Chest 0.39 (0.49*) Ð Ð Waist 0.48* (0.50**) 0.42 (0.77**) Hips 0.40* (0.35) 0.40 (0.50*) Thigh 0.47* (0.38) 0.40 (0.52*) Waist ± hip ratio 0.24 (0.32) 0.16 (0.52*) *P < 0.05; **P < adults to estimate intra-abdominal fat deposition, is not signi cantly associated in this group of 13-y-olds. This is also the case for waist ± thigh ratio. For the 13-y-old boys in this study, the regression analyses showed that the best eld indicators of intra-abdominal fat are truncal skinfolds. In boys, the horizontal abdominal skinfold explains 48% of variance (SEE ˆ 8.1 cm 2, P in ˆ 0.05). For girls the equivalent is the subscapular skinfold which explains 41% of variance in intra-abdominal fat area (SEE ˆ 7.8 cm 2, P in ˆ 0.05). This is encouraging because similar ndings were found with prepubertal children by Goran et al. 8 Discussion The combination of these two MRI studies provides some early insight into the complex nature of fat distribution change during the pubescent years. The data indicate a difference in patterns of change between boys and girls. There is some evidence that boys are beginning to deposit a greater percentage of their fat in intra-abdominal sites. Although girls have deposited signi cantly more fat overall than boys, it is distributed mainly in subcutaneous sites. The amount of intra-abdominal fat as a percentage of total adipose tissue at this age remains small for both boys and girls. Six of the girls and ve of the boys could be considered overweight as de ned by the 85th percentile of the UK age-related BMI percentiles of Only one girl and two boys were above the 95th percentile, which would class them as obese. Greater insight will be provided if more of these children show large increases in their overall levels of fatness as they grow into adulthood. Examination of patterns of change in adipose tissue deposition at the individual level reveals considerable heterogeneity within each sex. This would be consistent with genotype expression. The data also show how dif cult it is to nd standard eld estimators of total fat and intra-abdominal fat that are valid across maturational years. Although skinfolds still produced moderate associations with intra-abdominal fat, they did not show the same increases in body fatness indicated by underwater weighing and magnetic resonance imaging. Furthermore, different indicators of intra-abdominal fat emerged in study 2 when compared to study 1. For girls in study 1, all skinfolds were highly related to intra-abdominal fat but by study 2, with the exception of trunk skinfolds, these had reduced considerably to moderate or insigni cant levels. By age 13 therefore, abdominal fat is less closely related to general subcutaneous deposition. In boys, the reverse has occurred. In study 2 skinfold measures are more closely related to abdominal fat deposition than in study 1. Where boys deposit subcutaneous fat, this is re ected in increase intra-abdominal deposition. By

7 study 2 the best eld estimates of intra-abdominal fat for boys and girls are provided by trunk skinfolds. However, as with study 1, sensitivity remains inadequate for clinical classi cation. Waist ± hip ratio has been widely used to estimate intra-abdominal fat with adults. According to the evidence presented here and in study 1, albeit with a limited sample of white British children, WHR is invalid as a proxy measure of abdominal fat for peri-pubescent youngsters. In the case of adolescent girls, this was also stated by de Ridder et al. 11 However, its use continues to be reported. 20 Intraabdominal fat remains a small percentage of total fat and it is likely that it does not have suf cient impact on overall shape and is not re ected in waist circumference. In conclusion, the data provide initial evidence of the emergence of sex-speci c fat deposition patterns in early adolescence. However, there remains considerable variation in fat patterning within sexes. Magnetic resonance imaging continues to provide a valuable method for the study of the development of fat deposition in young people. References 1 Ashwell M, Cole TJ, Dixon AK. Obesity: a new insight into the anthropometric classi cation of fat distribution shown by computed tomography. Br Med J 1985; 301: 203 ± DespreÂs J-P, Moorjani S, Lupien PJ, Tremblay A, Nadeau A, Bouchard C. Regional distribution of body fat, plasma lipoproteins, and cardiovascular disease. Arteriosclerosis 1990; 10: 495 ± Brambilla P, Manzoni P, Sironi S, Simone P, Del Maschio A, Di Natale B, Chiumello G. Peripheral and abdominal adiposity in childhood obesity. Int J Obes Relat Metab Disord 1994; 18: 795 ± Caprio S, Hyman LD, McCarthy S, Lange R, Bronson M, Tamborlane WV. Fat distribution and cardiovascular risk factors in obese adolescent girls: importance of the intraabdominal fat depot. Am J Clin Nutr 1996; 64: 12 ± Ross R, Rissanen J, Pedwell H, Clifford J, Shragge P. In uence of diet and exercise on skeletal muscle visceral adipose tissue in men. J Appl Physiol 1996; 81: 2445 ± Goran MI, Kaskoun MC, Shuman WP. Intra-abdominal adipose tissue in young children. Int J Obes Relat Metab Disord 1995; 19: 279 ± Goran MI, Nagy TR, Treuth MT, Trowbridge C, Dezenberg C, McGloin A, Gower BA. Visceral fat in Caucasian and African- American pre-pubertal children. Am J Clin Nutr 1997; 65: 1703 ± Goran MI, Gower BA, Treuth M, Nagy TR. Prediction of intra-abdominal and subcutaneous abdominal adipose tissue in healthy prepubertal children. Int J Obes Relat Metab Disord 1998; 22: 549 ± Fox KR, Peters DM, Armstrong N, Sharpe P, Bell M. Abdominal fat deposition in 11-year-old children. Int J Obes Relat Metab Disord 1993; 17: 11 ± Peters DM, Fox KR, Armstrong N, Sharpe P, Bell M. Estimation of body fat and body fat distribution in 11-yearold children using magnetic resonance imaging and hydrostatic weighing, skinfolds and anthropometry. Am J Hum Biol 1994; 6: 237 ± De Ridder CM, de Boer RW, Seidell JC, Nieuwenhoff CM, Jeneson JAL, Bakker CJG, Zonderland ML, Erich WBM. Body fat distribution in pubertal girls quanti ed by magnetic resonance imaging. Int J Obes Relat Metab Disord 1992; 16: 443 ± Dooms GC, Hricak H, Margulis AR, de Greer G. MR imaging of fat. Radiology 1986; 158: 51 ± Ross R, Leger L, Morris D, de Guise J, Guardo R. Quanti cation of adipose tissue by MRI: relationship with anthropometric variables. J Appl Physiol 1992; 72: 787 ± Ross R, Shaw KD, Martel Y, de Guise J, Avruch L. Adipose tissue distribution measured by magnetic resonance imaging in obese women. Am J Clin Nutr 1993; 57: 470 ± Lohman TG, Roche AF, Martorell R (eds). Anthropometric Standards Reference Manual. Human Kinetics: Champaign, IL, Slaughter MH, Lohman TG, Boileau RA, Horswill CA, Stillman RJ, van Loan MD, Membden DA. Skinfold equations for estimation of body fatness in children and youth. Hum Biol 1988; 60: 709 ± Tanner JM. Growth at Adolescence, 2nd Edn. Blackwell Scienti c: Oxford, SPSS for Windows Release 9.0. SPSS Inc.: Chicago, IL, Cole TJ, Freeman JV, Preece MA. Body mass index reference curves for the UK, Arch Dis Child 1995; 73: 25 ± Gillum RF. Distribution of waist-to-hip ratio, other indices of body fat distribution and obesity and associations with HDL cholesterol in children and young adults aged 4 ± 19 y: the Third National Health and Nutrition Examination Survey. Int J Obes Relat Metab Disord 1999; 23: 556 ±