Quality control, accuracy, and prediction capacity of dual energy X-ray absorptiometry variables and data acquisition.

The acquisition of soft tissue measurements, fat (chemical) or adipose tissue (morphological) quantities, is essential in clinical research and nutritional status and its associated health risks. This has led to a proliferation of methods for the in-vivo determination of body composition. None of the indirect in-vivo approaches to estimate body adipose tissue has been validated against direct dissection data except for the skinfold.Since the development of DEXA as a measurement tool of bone density and mineral content for the detection of osteoporosis, it became a tool for measuring regional and whole body masses, lean tissue, and fat. A number of validation attempts have been made, however mostly against other indirect in-vivo techniques. The purpose of this study was to conduct an in-vitro validation and quality control of the data acquisition of DEXA using dissection as the criterion method. Fourteen porcine hind legs were scanned with DEXA, weighed in air and water and dissected into skin, adipose tissue, muscle and bone. Normal distribution, means, standard deviations, paired student t-test, Pearson correlation coefficients, interclass correlations test, and a Bland-Altman plot were used. The results show systematically good to excellent correlations between DEXA and dissection data acquisition (r(2)=0.75 to 0.99), but absolute indirect DEXA and direct dissection values are significantly different (p<0.05). DEXA overestimates total weight, lean mass, and fat free mass and underestimates both mass and % adipose tissue. DEXA provides erroneous values for bone density. Data produced by DEXA are morphological, not chemical values, as claimed by the manufacturer. This 'pilot' study indicates that a 'simple' combination of skin, adipose tissue, muscle, and bone from hind legs may give an incomplete picture of reality as one needs to measure viscera, connective tissue, and air pockets of the body. Validation studies with intact bodies are advised.