@article{oai:kanazawa-u.repo.nii.ac.jp:00000504,
 author = {Demura, Shinichi and Sato, S. and Noguchi, T. and Nakata, Y.},
 issue = {1},
 journal = {Sport Sciences for Health},
 month = {May},
 note = {This study estimated the area of visceral fat at the L4-L5 level (VFAL4-5) measured by computed tomography (CT) from anthropometric and segmental percent fat variables. Subjects were 73 adults (50 men and 23 women) aged 24-78 years. Cross-validation was permormed with another 38 adults (25 men and 13 women) aged 21-80 years. The anthropometric variables examined were height, weight, waist circumference, hip circumference, sagittal diameter, and subcutaneous fat thickness (SFT) at 14 sites. SFT and segmental percent fat were measured by ultrasonography and dualenergy X-ray absorptiometry (DXA), respectively. A combination of suitable predictors of VFAL4-5 was derived by stepwise multiple regression analysis using these variables. A prediction equation was obtained that used seven predictors: sagittal diameter, waist circumference, three subcutaneous thickness variables (subscapula, chest 1 and abdomen), and segmental percent fat at the trunk and left leg (R=0.902, R2 =0.813, SEE=17.5 cm2). In a Bland-Altman procedure, systematic error was not found in the original group but was only found in women in the cross-validation group. The percentage of the SEE of the prediction equation for the mean VFAL4-5 value was 22.5% in the original group and 20.1% in the cross-validation group. Furthermore, the percentages of SD values of the error for the mean VFAL4-5 value were 21.1% in the original group and 22.2% in the cross-validation group. These values were comparable or superior to those in previous studies. This study provides a useful prediction equation for VFAL4-5 from anthropometry and segmental body composition variables. © Springer-Verlag Italia 2007., 金沢大学教育学部保健体育},
 pages = {16--22},
 title = {Prediction of visceral fat area from anthropometric and segmental body composition variables using computed tomography},
 volume = {2},
 year = {2007}
}