International Journal of Obesity (1998) 22, 655±660 ß 1998 Stockton Press All rights reserved 0307±0565/98 $12.00 http://www.stockton-press.co.uk/ijo Sagittal abdominal diameter as a practical predictor of visceral fat M Zamboni 1, E Turcato 1, F Armellini 1, HS Kahn 2, A Zivelonghi 1, H Santana 1, IA Bergamo-Andreis 3 and O Bosello 1 Institutes of 1 Geriatric Medicine and 3 Radiology, University of Verona, Policlinico of Borgo Roma, Verona, Italy, and 2 Department of Epidemiology and Surveillance Research, American Cancer Society, Atlanta, GA, USA OBJECTIVE: To evaluate the relationships between the supine sagittal abdominal diameter (SAD) and visceral fat, as well as to evaluate intra- and inter-observer reliability of sagittal diameter measurement. PIENTS: Twenty-eight women ranging in age from 27±78 y with a body mass index (BMI) ranging from 16.9±48.1 kg/m 2 and 23 men ranging in age from 32±75 y with BMI ranging from 20±41.6 kg/m 2. MEASUREMENT: Body fat distribution was measured by waist circumference, waist to hip ratio (WHR), SAD, anthropometrically assessed and a single slice of computed tomography (CT) at the L4-L5 level. RESULTS: In both genders, a signi cant association was found between visceral adipose tissue () and SAD, as evaluated by CT (women r ˆ 0.80; men r ˆ 0.83, P < 0.001), and SAD by anthropometry (women r ˆ 0.76; men r ˆ 0.82, P < 0.001), as well as between visceral and waist circumference (women r ˆ 0.76, men r ˆ 0.86, P < 0.001) and WHR (women r ˆ 0.57, P < 0.01, men r ˆ 0.80, P < 0.001). A signi cant association was also found between subcutaneous and SAD by anthropometry (women r ˆ 0.79, men r ˆ 0.74, P < 0.001). After adjusting for BMI, the association between subcutaneous and SAD was no longer signi cant in men and only moderately signi cant in women (r ˆ 0.42, P < 0.05), while the association between visceral and SAD by anthropometry remained signi cant in both genders (women r ˆ 0.63, P < 0.001; men r ˆ 0.66, P < 0.001). When the subjects were divided into two groups according to BMI (lean to moderately overweight women with BMI < 28 and men with BMI < 30 and obese women with BMI > 28 and men with BMI > 30) we found that the relationships between SAD by anthropometry, as well as SAD by CT and visceral, were higher in lean to moderately overweight subjects than in those who were obese. High inter-observer correlation was found concerning SAD measurement (r ˆ 0.99, P < 0.001). Intra- and interobserver precision as evaluated by coef cient of variation and intraclass correlation coef cient for SAD measurement was very high. CONCLUSION: Our study shows the usefulness of SAD by anthropometry to predict visceral fat and its very high inter- and intra-observer precision. Keywords: body fat distribution; sagittal diameter; waist and visceral fat Introduction There is growing evidence that regional body fat distribution, rather than body fat quantity, is a risk factor for cardiovascular disease morbidity and mortality. 1,2 For this reason the evaluation of regional body fat distribution seems to be clinically relevant. Computerised topography (CT) and magnetic resonance imaging (MRI) are the gold standard methods to evaluate body fat distribution and are the only methods that clearly differentiate between visceral and subcutaneous adipose tissue (). 3 Imaging techniques such as MRI and CT are costly, not easily accessible and thus not useful for epidemiological studies. 3 To overcome these limitations, some investigators have experimented with new methods such as ultrasound, to evaluate regional body fat distribution, 4 or attempted to determine which anthropometric measurement is the best correlate of visceral accumulation. 3 Waist girth, waist-to-hip ratio (WHR) Correspondence: Dr Mauro Zamboni, Institute of Geriatric Medicine, Ospedale Maggiore, Piazzale Stefani 1, 37126 Verona, Italy. Received 31 October 1997; revised 18 February 1998; accepted 3 March 1998 and supine sagittal abdominal diameter (SAD) are commonly used to predict visceral accumulation. 3±12 It has been shown that waist girth and SAD show closer association with visceral abdominal accumulation than WHR. 5,6 A few studies evaluated the relationship between visceral and SAD anthropometrically determined, 7±9 while the majority used SAD derived by CT scan. 4±6,9±12 Clinical and epidemiological studies need a lowcost, non-invasive method for the assessment of cardiovascular risk. Before using SAD anthropometrically assessed for clinical and epidemiological studies, more data are needed about this inter- and intra-observer precision. The aims of our study were to evaluate the association between SAD anthropometrically assessed and visceral as evaluated by CT, as well as to evaluate intra- and inter-observer reliability. Methods The study was conducted on 28 women ranging in age from 27±78 y with a body mass index (BMI) ranging
656 from 16.9± 48.1 kg/m 2 and 23 men ranging in age from 32±75 y with BMI ranging from 20 ± 41.6 kg/m 2. All the subjects were volunteers and gave their consent to the study, whose protocol was approved by our Ethical Committee. The following anthropometric variables were evaluated in all subjects: weight, height, BMI, waist and hip circumferences and SAD. Body weight was measured to the nearest 0.2 kg (Salus scale, Milan, Italy) and height to the nearest 0.5 cm using a stadiometer (Salus). BMI, an index of body fatness, was calculated as body weight adjusted for stature (kg/m 2 ). Circumferences were measured to the nearest 0.5 cm using a 1 cm wide metal measuring tape. The waist circumference was measured with subjects standing at the minimum abdominal circumferences, between the xyphoid process and the umbilicus. The hip circumference was measured at the outermost points on the greater trochanters. WHR is the ratio of these two circumferences. SAD was measured with a portable, sliding-beam, abdominal caliper, while the subjects were in a supine position (Robbia., Verona, Italy); this instrument was hand-made under our supervision by an artisan. The SAD was measured at the largest supine anteroposterior diameter between xyphoid process and umbilicus. The subjects were asked to inhale and exhale gently, and the arm of the caliper was brought down to touch the abdominal wall without compression. All subjects were examined on a standard at hospital bed. The measurements of SAD were taken in two positions with hips fully extended or exed. In 14 subjects duplicate measurements were taken. In 18 of the subjects (six men ranging in age from 36 ±74 y and ranging in BMI from 20 ±38 and 12 women ranging in age from 32±78 y and ranging in BMI from 16.9 ± 43.7) the SAD was measured by two different operators, using both the hip positions in order to evaluate reproducibility. Each operator was unaware of the results obtained by the other and the two measurements were taken on the same day. CT measurements Body fat distribution was determined by CT (Siemens Instrument, Erlangen, Germany) according to SjoÈstroÈm. 13 Total abdominal, visceral abdominal, subcutaneous abdominal and visceral to subcutaneous abdominal ratio were evaluated by a single scan at L4-L5 level. A lateral topogram (scout view) was performed to identify the standard level with precision. The usual radiographic parameters for abdominal investigations were used: 125 kv, 350 ma; scanning time, 4 s; slice thickness, 8 mm. Regions of interest were outlined using a light pen cursor (joystick). The preselected attenuation interval chosen for fat was 7150 to 750 Houns eld units. 14,15 This range was chosen after performing quality checks on our CT machine with a water phantom. Daily controls must refer to these quality checks, usually performed by a Siemens technician for comparison. The reproducibility of the method (coef- cient of variation) was 1.1%. Statistical methods Results are presented as the mean standard deviation (s.d.). To test the associations between different variables, simple and partial correlation analyses were performed; slopes and intercepts of the regression lines were compared. Differences between the two observers were evaluated by paired t test. To test intra- and inter-observer precision, the coef cient of variation (CV) and intraclass correlation coef cient (ICC) were computed. CV was computed by dividing the square root of the within-subject variance by the overall mean value of the measurement. ICC was the ratio of between-subject variance to the sum of the between and within-subject variance. 16 A two-tailed signi cance level of 0.05 was used throughout. Results Characteristics of the study sample are shown in Table 1. In Table 2, the correlation matrix for SAD (as evaluated by both CT and caliper) in relation to the other anthropometric variables in women are reported. A signi cant association was found between visceral and SAD as evaluated by CT (r ˆ 0.80, P < 0.001), SAD by anthropometry assessed with hips extended (r ˆ 0.76, P < 0.001) and with hips exed (r ˆ 0.76, P < 0.001). A signi cant association was also found between visceral and waist circumference (r ˆ 0.76, P < 0.001) and WHR (r ˆ 0.57, P < 0.01). A signi cant association was also found between subcutaneous and SAD by anthropometry assessed with hips extended (r ˆ 0.79, P < 0.001) and with hips exed (r ˆ 0.78, P < 0.001). A signi cant Table 1 Characteristics of study sample (mean s.d.) Women (n ˆ 28) Men (n ˆ 23) P Age (y) 53.7 17.1 57.1 16.9 NS Body weight (kg) 82.5 24.4 83.6 21.7 NS Height (cm) 159.9 6.7 169.5 5.9 0.001 BMI 31.3 8.6 29.0 6.5 NS Waist (cm) 94.3 17.7 99.2 12.9 NS Hip (cm) 110.2 17.9 101.5 9.8 0.05 WHR 0.85 0.07 0.97 0.06 0.001 SAD CT (cm) 24.5 5.6 23.7 4.7 NS SAD-E (cm) 23.5 4.7 24.0 4.4 NS SAD-F (cm) 23.6 4.8 24.1 4.5 NS Total (cm 2 ) 491.7 252.5 402.3 218.0 NS Subcutaneous (cm 2 ) 363.2 202.5 250.9 189.5 0.05 Visceral (cm 2 ) 141.3 86.3 167.7 83.2 NS BMI ˆ body mass index; WHR ˆ waist to hip ratio; SAD ˆ sagittal diameter; CT ˆ computed tomography; SAD-E ˆ sagittal diameter with hips extended; SAD-F ˆ sagittal diameter with hips exed; ˆ adipose tissue; NS ˆ not statistically signi cant.
Table 2 Correlation matrix for age and anthropometric variables in women (n ˆ 28) 657 Age Body weight BMI Waist Hip WHR SAD CT SAD-E SAD-F Total Subcutaneous Visceral Age 1 Body weight 70.45* 1 BMI 70.35* 0.90*** 1 Waist 70.28 0.84*** 0.75*** 1 Hip 70.39* 0.90*** 0.76*** 0.88*** 1 WHR 0.12 0.14 0.17 0.51** 0.04 1 SAD CT 70.32 0.82*** 0.77*** 0.94*** 0.88*** 0.39* 1 SAD-E 70.30 0.90*** 0.84*** 0.97*** 0.90*** 0.39* 0.96*** 1 SAD-F 70.30 0.90*** 0.83*** 0.97*** 0.91*** 0.38* 0.96*** 1 Total 70.21 0.78*** 0.75*** 0.83*** 0.79*** 0.30 0.82*** 0.84*** 0.83*** 1 Subcutaneous 70.36 0.80*** 0.75*** 0.77*** 0.80*** 0.17 0.76*** 0.79*** 0.78*** 0.96*** 1 Visceral 0.11 0.53** 0.57** 0.76*** 0.56** 0.57** 0.80*** 0.76*** 0.76*** 0.70*** 0.52** 1 BMI ˆ body mass index; WHR ˆ waist to hip ratio; SAD ˆ sagittal diameter; CT ˆ computed tomography; SAD-E ˆ sagittal diameter hips extended; SAD-F ˆ sagittal diameter hips exed; ˆ adipose tissue. * P < 0.05; ** P < 0.01; *** P < 0.001 association was found between SAD as evaluated by CT and SAD by anthropometry assessed with hips extended (r ˆ 0.96, P < 0.001) and with hips exed (r ˆ 0.96, P < 0.001). In Table 3, the correlation matrix for SAD (as evaluated by both CT and caliper) in relation to the other anthropometric variables in men are reported. A signi cant association was found between visceral and SAD as evaluated by CT (r ˆ 0.83, P < 0.001), SAD by anthropometry assessed with hips extended (r ˆ 0.82, P < 0.001) and with hips exed (r ˆ 0.82, P < 0.001). A signi cant accociation was also found between visceral and waist circumference (r ˆ 0.86, P < 0.001) and WHR (r ˆ 0.80, P < 0.001). A signi cant association was also found between subcutaneous and SAD by anthropometry assessed with hips extended (r ˆ 0.74, P < 0.001) and with hips exed (r ˆ 0.76, P < 0.001). A signi cant association was found between SAD as evaluated by CT and SAD by anthropometry assessed with hips extended (r ˆ 0.97, P < 0.001) and with hips exed (r ˆ 0.97, P < 0.001). The relationships between visceral and SAD by anthropometry assessed with hips extended and exed in men and women are shown in Figure 1. Analysis of slopes and intercepts did not reveal any signi cant differences between women and men. As a consequence only one regression line was drawn. After adjusting for BMI in both genders, the association between subcutaneous and sagittal diameter was signi cant in women (r ˆ 0.42, P < 0.05) but not in men (r ˆ 0.03, not statistically signi cant (NS)), while the association between visceral and SAD anthropometrically assessed with hips extended and exed, was still signi cant in both genders (women r ˆ 0.63 and r ˆ 0.63 respectively, P < 0.001; men r ˆ 0.66 and r ˆ 0.70 respectively, P < 0.001) (data not shown in table). After adjusting for BMI, in both genders, the associations between visceral and waist and WHR were still signi cant (women r ˆ 0.61 and r ˆ 0.58 respectively, P < 0.01; men r ˆ 0.81 and r ˆ 0.68 respectively, P < 0.001) (data not shown in table). We performed correlation analyses for SAD (as evaluated by both CT and caliper), as well as for waist and for WHR, with visceral and subcutaneous after subdividing our sample into two groups on the basis of BMI: lean to moderately overweight (women with BMI < 28 and men with BMI < 30), and obese (women with BMI > 28 and men with BMI< 30) (Table 4); women and men were pooled together in Table 3 Correlation matrix for age and anthropometric variables in men (n ˆ 23) Age Body weight BMI Waist Hip WHR SAD CT SAD-E SAD-F Total Subcutaneous Visceral Age 1 Body weight 70.57** 1 BMI 70.45* 0.95*** 1 Waist 70.40 0.95*** 0.92*** 1 Hip 70.45* 0.95*** 0.95*** 0.92*** 1 WHR 0.14 0.64** 0.57** 0.79*** 0.50* 1 SAD CT 70.55** 0.96*** 0.92*** 0.96*** 0.91*** 0.72*** 1 SAD-E 70.50* 0.96*** 0.91*** 0.98*** 0.93*** 0.73*** 0.97*** 1 SAD-F 70.50* 0.97*** 0.93*** 0.98*** 0.93*** 0.73*** 0.97*** 1*** 1 Total 70.45* 0.89*** 0.96*** 0.86*** 0.94*** 0.43* 0.88*** 0.85*** 0.86*** 1 Subcutaneous 70.57** 0.83*** 0.87*** 0.72*** 0.89*** 0.23 0.78*** 0.74*** 0.76*** 0.93*** 1 Visceral 70.21 0.75*** 0.68*** 0.86*** 0.70*** 0.80*** 0.83*** 0.82*** 0.82*** 0.60** 0.38 1 BMI ˆ body mass index; WHR ˆ waist to hip ratio; SAD ˆ sagittal diameter; CT ˆ computed tomography; SAD-E ˆ sagittal diameter hips extended; SAD-F ˆ sagittal diameter hips exed; ˆ adipose tissue. * P < 0.05; ** P < 0.01; *** P < 0.001.
658 Figure 1 Relationships between visceral adipose tissue () and sagittal abdominal diameter (SAD) by anthropometry, assessed with hips extended and exed in men and women. Analysis of slopes and intercepts did not reveal any signi cant differences between women and men. As a consequence only one regression line was drawn. this analysis because the regression lines were similar in the two genders. We found that the relationships between SAD by anthropometry, as well as SAD by CT, and visceral were higher in lean to moderately overweight subjects than in the obese. Figure 2 shows the relation between SAD measured by two observers in 18 subjects. High inter-observer Figure 2 Relationship between sagittal abdominal diameter (SAD) measured by two observers of 18 subjects. correlation was found, using both positions to anthropometrically determine SAD (with hips extended: r ˆ 0.99, P < 0.001; with hips exed: r ˆ 0.99, P < 0.001). Table 4 Correlation analyses for visceral adipose tissue (), subcutaneous and anthropometric variables after subdividing all subjects into two groups on the basis of body mass index (BMI) Lean to moderately overweight subjects (n ˆ 23) Obese subjects (n ˆ 28) Subcutaneous Visceral Subcutaneous Visceral r r r r Waist 0.29 0.87*** 0.31 0.43* WHR 0.03 0.71*** 0.33 0.49** SAD CT 0.40 0.79*** 0.50** 0.51** SAD-E 0.38 0.86*** 0.40* 0.43* SAD-F 0.37 0.86*** 0.40* 0.41* Total 0.94*** 0.81*** 0.89*** 0.04 Lean to moderately overweight subjects: women with BMI < 28 (n ˆ 13) and men with BMI < 30 (n ˆ 10). Obese subjects: women with BMI > 28 (n ˆ 15) and men with BMI > 30 (n ˆ 13). WHR ˆ waist to hip ratio; SAD ˆ sagittal diameter; CT ˆ computed tomography; SAD-E ˆ sagittal diameter hips extended; SAD-F ˆ sagittal diameter hips exed. * P < 0.05; ** P < 0.01; *** P < 0.001.
Table 5 Paired t test for the differences in sagittal diameter between the two operators in 18 subjects (mean s.d.) Paired t test for the differences in SAD between the two operators are reported in Table 5. The two operators obtained almost identical values of SAD assessed in both the positions. Intra- and inter-observer precision was very high (intra-observer CV 0.58%, ICC 99.9% with hips extended; intra-observer CV 0.57%, ICC 99.9% with hips exed; inter-observer CV ˆ 1.5%, ICC 99.6% with hips extended; inter-observer CV ˆ 0.9%, ICC 99.8% with hips exed) (data not shown in table). Discussion Operator1 Operator 2 P SAD-E 25.4 4.2 24.8 4.5 0.83 SAD-F 25.5 4.3 24.8 4.6 0.94 SAD-E: sagittal diameter hips extended; SAD-F ˆ sagittal diameter hips exed. Our data show a strong relationship between SAD anthropometrically assessed and visceral area as evaluated by CT, even after adjusting for BMI. Our study also shows very high precision of the SAD anthropometrically assessed. SAD has been previously suggested as a good index for estimating visceral, 3±12 but most papers evaluated the relationships between SAD and visceral area both derived by CT. 4±6,9±12 It is well known that the gold standard to evaluate regional body fat distribution, would be multi-slice CT for determination of the volume of visceral, rather than single-slice CT for determination of the area of visceral. 3 However, a very high correlation has been found between the visceral area, taken by a single slice of CT at L4-L5 level (as in our study) and the volume of. 3 It has been found that correlation coef cients between visceral area and SAD, derived by the same CT, ranged from 0.46 ± 0.96, with poorer association in obese people. 3,10 In agreement with these studies 3,10 we found that SAD, both by anthropometry and by CT, is a better predictor of visceral in lean to moderately overweight subjects than in the obese. The different degree of association between SAD and visceral in lean to moderately overweight subjects and in obese subjects, may result from the inability of SAD by itself to distinguish between abdominal visceral and subcutaneous. Our results showing quite similar correlation between SAD and visceral and subcutaneous areas are in line with previous studies. 7,9,10 These ndings do not support the hypothesis that visceral fat, not subcutaneous fat, increases the depth of the abdomen. 5 Among our subjects, however, once we adjusted for BMI, the association of visceral with SAD remained signi cant, but in men, the association with subcutaneous did not. This observation seems to support the usefulness of SAD as an index of visceral adipose tissue distribution and is in line with our previous ndings obtained after correcting SAD derived by CT, by subtracting the ultrasound subcutaneous thickness. 17 Our study showing very signi cant association between SAD by CT and SAD by anthropometry, con rms previous studies 7,9 and shows the possibility of using SAD by anthropometry instead of that by CT as an index of visceral. Some clinical studies have already supported the usefulness of SAD, both by anthropometry and by CT, as a good marker of visceral adipose tissue distribution. 18±22 SAD by anthropometry has been shown to be related to cardiovascular risk factors. 18 Recently, Kahn et al 19 reported that SAD by anthropometry was better than waist circumference for discriminating men with ischaemic heart disease from control subjects. This nding complements the observation that men with coronary artery disease have greater abdominal visceral than controls, but not greater abdominal subcutaneous. 23 A negative relationship has also been shown between SAD anthropometrically assessed and physical activity, 21 and the seven-year changes in visceral are better predicted by changes in SAD by CT, than by changes in WHR in women. 22 Recently, Pouliot et al 11 observed that in 81 men and 70 women with age ranging from 23±50 y, both waist and SAD by CT were strongly related to abdominal visceral and that waist circumference values above 100 cm and SAD by CT above 25 cm are associated with potentially atherogenic metabolic abnormalities. The degree of correlation between waist circumference and SAD with visceral area in our study, seems to be quite similar both before and after adjusting for BMI. Previous studies, but not all, showed that SAD was superior to waist, in predicting the visceral. 5,7,9 This discrepancy may have resulted from methodological differences, with the studies showing that waist circumference was superior to SAD in predicting visceral, using a single-scan CT rather than volume determinations from multiple scans. 18 It is possible that our results may have underestimated the utility of the SAD, because we evaluated visceral with a single-ct scan, rather than a multi-scan estimate of visceral volume. Our choice to measure SAD as the largest supine antero-posterior diameter, may have also underestimated the utility of SAD as predictor of visceral ; in fact others 18,19 have suggested the measurement of SAD at the iliac crest as the best predictor of visceral ; eventually, further studies will be needed to compare the relationships between visceral and 659
660 SAD anthropometrically assessed at the iliac crest and at the largest supine antero-posterior diameter. Our experience with measurements obtained by different observers, supports a previous report that SAD measurements can be obtained with a high degree of precision. 24 Our study population covered a wider range of age and BMI than previous reports, 3,24 although we have no data with which to compare the precision of measuring supine waist circumference or standing SAD by anthropometry. Conclusion Our data show that SAD obtained by anthropometry in supine subjects is a good index of visceral adipose tissue distribution. Epidemiological and clinical studies may take advantage of this easy, cheap and highly precise method to evaluate visceral adipose tissue distribution. Acknowledgements The study was supported by grants from Regione Veneto, Giunta Regionale, Ricerca Sanitaria Finalizzata, 730/01/96, Venezia, Italia. References 1 Bouchard C, Bray GA, Hubbard VS. Basic and clinical aspects of regional fat distribution. Am J Clin Nutr 1990; 52: 946±950. 2 Zamboni M, Armellini F, Cominacini L, Turcato E, Todesco T, Bissoli L, Micciolo R, Bergamo-Andreis IA, Bosello O. Obesity and regional body-fat distribution in men: separate and joint relationships to glucose tolerance and plasma lipoproteins. Am J Clin Nutr 1994; 60: 682±687. 3 van der Kooy K, Seidell JC. Techniques for the measurement of visceral fat: a practical guide. Int J Obes 1993; 17: 187±196. 4 Armellini F, Zamboni M, Castelli S, Micciolo R, Mino A, Turcato E, Rigo L, Bergao-Andreis IA, Bosello O. Measured and predicted total and visceral adipose tissue in women. Correlations with metabolic parameters. Int J Obes 1994; 18: 641±647. 5 Kvist H, Chowdhury B, GrangaÈrd U, TyleÂn U, SjoÈstroÈm L. Total and visceral adipose-tissue volumes derived from measurements with computed tomography in adult men and women: predictive equations. Am J Clin Nutr 1988; 48: 1351±1361. 6 Lemieux S, Prud'homme D, Bouchard C, Tremblay A, DespreÁs JP. A single threshold value of waist girth identi es normal-weight subjects with excess visceral adipose tissue. Am J Clin Nutr 1996; 64: 685±693. 7 Svendsen OL, Hassager C, Bergmann I, Christiansen C. Measurement of abdominal and intra-abdominal fat in postmenopausal women by dual energy X-ray absorptiometry and anthropometry: comparison with computerized tomography. Int J Obes 1993; 17: 45±51. 8 Jensen MD, Kanaley JA, Reed JE, Sheedy PF. Measurement of abdominal and visceral fat with computed tomography and dual-energy X-ray absorptiometry. Am J Clin Nutr 1995; 61: 274±278. 9 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. Brit J Nutr 1993; 70: 47±58. 10 DespreÂs JP, Prud'homme D, Pouliot MC, Tremblay A, Bouchard C. Estimation of deep abdominal adipose-tissue accumulation from simple anthropometric measurements in men. Am J Clin Nutr 1991; 54: 471±477. 11 Pouliot MC, DespreÁs JP, Lemieux S, Moorjani S, Bouchard C, Tremblay A, Nadeau A, Lupien PJ. Waist circumference and abdominal sagittal diameter: best simple anthropometric indexes of abdominal visceral adipose tissue accumulation and related cardiovascular risk in men and women. Am J Cardiol 1994; 73: 460±468. 12 Lemieux S, Prud'homme D, Bouchard C, Tremblay A, DespreÁs JP. Sex differences in the relation of visceral adipose tissue accumulation to total body fatness. Am J Clin Nutr 1993; 58: 463±467. 13 SjoÈstroÈm L. New aspects of weight-for-height indices and adipose tissue distribution in relation to cardiovascular risk and total adipose tissue volume. In: Berry EM, Blondheim SH, Elihau HE, Shafrir E (eds). Recent Advances in Obesity Research: V. John Libbey: London, 1987, pp 66±76. 14 Ashwell M, Cole TJ, Dixon AK. Obesity: new insight into the anthropometric classi cation of fat distribution shown by computed tomography. BMJ 1985; 290: 1692±1694. 15 Seidell JC, Oosterlee A, Thijssen MAO, Burema J, Deurenberg P, Hautvast JGAJ, Ruijs JHJ. Assessment of intraabdominal and subcutaneous abdominal fat: relation between anthropometry and computed tomography. Am J Clin Nutr 1987; 45: 7±13. 16 Mueller SH, Martorell R. Reliability and accuracy of measurement. In: Lohman TB, Roche AF, Martorell R (eds). Anthropometric Standardization Reference Manual: Abridged Edition, Human Kinetics. Champaign, IL, 1988, pp 83±87. 17 Armellini F, Zamboni M, Harris T, Micciolo R, Bosello O. Sagittal diameter minus subcutaneous thickness. An easy-toobtain parameter that improves visceral fat prediction. Obes Res 1997; 5: 315±320. 18 SjoÈstroÈm CD, HaÈkangaÈrd AC, Lissner L, SjoÈstroÈm L. Body compartment and subcutaneous adipose tissue distributionð risk factor patterns in obese subjects. Obes Res 1995; 3: 9±22. 19 Kahn HS, Austin H, Williamson DF, Arensberg D. Simple anthropometric indices associated with ischemic heart disease. J Clin Epidemiol 1996; 49: 1017±1024. 20 Seidell JC, Andres R, Sorkin JD, Muller DC. The sagittal waist diameter and mortality in men: the Baltimore Longitudinal Study on Aging. Int J Obes 1994; 18: 61±67. 21 Houmard JA, McCulley C, Roy LK, Bruner RK, McCammon MR, Israel RG. Effects of exercise training on absolute and relative measurements of regional adiposity. Int J Obes 1994; 18: 243±248. 22 Lemieux S, Prud'homme D, Tremblay A, Bouchard C, DespreÁs JP. Anthropometric correlates to changes in visceral adipose tissue over seven years in women. Int J Obes 1996; 20: 618±624. 23 Zamboni M, Armellini F, Sheiban I, De Marchi M, Todesco T, Bergamo-Andreis IA, Cominacini L, Bosello O. Relation of body fat distribution in men and degree of coronary narrowings in coronary artery disease. Am J Cardiol 1992; 70: 1135±1138. 24 Williamson DF, Kahn HS, Worthman CM, Burnette JC, Russell CM. Precision of recumbent anthropometry. Am J Hum Biol 1993; 5: 159±167.