Adipose tissue volume determination in males by computed tomography and 40K

Seventeen healthy male volunteers with weights ranging from 54 to 145 kg were examined with a Philips Tomoscan 310. The upper attenuation limit of adipose tissue was determined to be -30 HU. The lower attenuation limit was set to -190 HU. Regional and total adipose tissue volumes were calculated from the adipose tissue areas of 22 scans and from the distances between these scans. Three different mathematical formulas were used, which all gave similar results. The adipose tissue area of several trunk scans, as well as the elbow, showed very high correlations (r greater than 0.96) versus the volume determinations based on 22 scans. The visceral adipose tissue area of scan L2-L3 showed a higher correlation (r = 0.986) than any other single scan versus the visceral adipose tissue volume. Total adipose tissue volume determinations with ten selected scans correlated very closely with the results obtained from 22 scans (r = 0.997). The adipose tissue volume of the head and neck region was 1.9 +/- 1.0 per cent of the total volume. Corresponding figures for other regions were: arms 6.8 +/- 1.0 per cent, legs 29.0 +/- 7.3 per cent, subcutaneous part of the trunk 41.4 +/- 7.4 per cent and the visceral region 20.9 +/- 7.0 per cent. With greater total adipose tissue volumes the percentage of the subcutaneous adipose tissue of the trunk increased (r = 0.686; P less than 0.005). There was a very strong negative relationship between the fractional amount of adipose tissue in the legs and in the trunk (r = 0.993, P less than 0.001). The potassium contents of fat-free mass and lean body mass were deduced to be 64.7 and 71.0 mmol/kg, respectively. These calculations were based on adipose tissue volume determinations by computed tomography, on 40K measurements and on the assumption that the volume proportions of fat, water and protein in adipose tissue were 85:13.7:1.3. By using computed tomography (CT) as a standard an optimal weight (W) for height (H) index was constructed by using an iterative correlation technique. The optimal index, i.e. highest correlation and lowest error versus ATCT was found for W/H0.9.