Relationships between computed tomography tissue areas, thicknesses and total body composition

Abstract

Computed tomography (CT) scanning was evaluated for its potential application to body-composition research. Three cross-sections (upper leg, abdomen, chest including upper arms) were scanned in 41 healthy men (mean age 57.6 years). Subcutaneous fat thicknesses measured at specific sites on the CT scans were correlated with the total area of fat from the same scans. For the chest and leg cross-sections, correlations were highly significant. Subcutaneous fat thicknesses at the abdomen were relatively poorer correlates of total abdomen fat area, because they were unrelated to intra-abdominal fat. Correlation analyses were performed between fat areas of each cross-section and total fat weight (by 40K counting), and the abdomen yielded the highest correlations. Multiple regression was used to predict abdomen fat area from external anthropometry, and abdomen circumference plus one skinfold provided excellent prediction of total abdomen fat area (R2 = 0.79). Subcutaneous or intra-abdominal fat areas separately were not predicted as well by external measurements. When lean body weight was predicted by multiple regression, leg lean area was the best predictor of any anatomical cross-section.