P A Tataranni1, D J Pettitt, E Ravussin. 1. Clinical Diabetes and Nutrition Section, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Phoenix, AZ 85016, USA.
Abstract
OBJECTIVE: Dual-energy X-ray absorptiometry (DEXA) is rapidly becoming the method of choice for body composition measurements. We tested inter-machine variability because large differences in body composition measurements between different DEXA machines have recently been reported. DESIGN: Comparison of total body scans using 2 DEXA machines from the same manufacturer (DPX-L; Lunar Co, Madison, WI) on 10 volunteers (5M/5F). RESULTS: We observed statistically significant differences between the 2 machines for all mean values of body composition variables as a result of a systematic underestimation of bone mineral and overestimation of fat tissue by one machine vs the other. However, the magnitude of the observed differences was small (namely bone mineral +68 +/- 57 g; percent body fat -1.7 +/- 1%, Mean +/- s.d.). CONCLUSION: Differences do exist in the performances of 2 DEXA machines from the same manufacturer. Although the differences reported in the present study are small, emphasis should be given in pre-testing machines when multiple apparatuses are used in a study. Also, because the observed error was systematic, randomized designs are necessary when more than one DEXA machine is used in longitudinal/intervention study. Better yet, manufacturers of DEXA machine should standardize their equipment to ensure the best consistency between machines.
OBJECTIVE: Dual-energy X-ray absorptiometry (DEXA) is rapidly becoming the method of choice for body composition measurements. We tested inter-machine variability because large differences in body composition measurements between different DEXA machines have recently been reported. DESIGN: Comparison of total body scans using 2 DEXA machines from the same manufacturer (DPX-L; Lunar Co, Madison, WI) on 10 volunteers (5M/5F). RESULTS: We observed statistically significant differences between the 2 machines for all mean values of body composition variables as a result of a systematic underestimation of bone mineral and overestimation of fat tissue by one machine vs the other. However, the magnitude of the observed differences was small (namely bone mineral +68 +/- 57 g; percent body fat -1.7 +/- 1%, Mean +/- s.d.). CONCLUSION: Differences do exist in the performances of 2 DEXA machines from the same manufacturer. Although the differences reported in the present study are small, emphasis should be given in pre-testing machines when multiple apparatuses are used in a study. Also, because the observed error was systematic, randomized designs are necessary when more than one DEXA machine is used in longitudinal/intervention study. Better yet, manufacturers of DEXA machine should standardize their equipment to ensure the best consistency between machines.
Authors: Bruce W Bailey; Ciera L Bartholomew; Caleb Summerhays; Landon Deru; Sharla Compton; Larry A Tucker; James D LeCheminant; Joseph Hicks Journal: J Obes Date: 2019-12-01