Page 118 - Go4it

Chapter 7
than others, but all lack accuracy. The Gray equations performed best. Our new equation predicted FFM in 76% of the obese adolescents in our study group. However, the equation awaits external validation. There is no consensus regarding the level of error that is acceptable when measuring FFM. In theory the 2.5% cut-off value is clinically more appropriate than the 5% cut-off value, since an error of about 3 kg (see Table 2: 5% of FFM-DXA 56.1 kg =3kg) appears quite large. In case a cut-off of 2.5% was used, the maximum accuracy level found was 35%. However, in this study the cut-off level was mainly used to rank the available FFM-BIA equations from good to poor.
There is a whole range of published FFM-BIA equations, although only 3 originally developed in obese adolescents (8-10). However, all 3 failed to produce acceptable FFM values comparable to DXA when applied to our sample of obese adolescents. Therefore, in this study, other FFM-BIA equations were considered, both based on a larger range of children with respect to age as well as weight and BMIsds. In an earlier study on energy expenditure we showed that obese adolescents have such stature and body mass that equations relevant for energy expenditure should rather be based on the 18+ category and not the normal 12-18 year age range (24). In fact, the Gray equation is developed in adults (25m, 62f), with BMI varying from 19.6 tot 53.3 kg/m2 and percent body fat from 8.8 tot 59%. Compared to DXA in our study, the BIA algorithm that is part of the devices overestimated FFM in obese children and adolescents (9,25). In healthy adult persons, the assumption is that 73.2% of the lean body mass consists of total body water (26). Wells et al. found that in children (aged 4-23 years) the FFM hydration is higher (mean 75%), and differed by age and sex (27). In obese, the hydration of the lean body mass is also great er than 73.2%. This increase in hydration in obese compared with non-obese individuals averaged 1% and reached 2% in extreme obesity (28). So both, childhood and obesity could cause an overestimation of FFM, which in turn could underestimate FM. As far as we known, all selected developed FFM equations (see Table 2) are based on the assumption of the hydration factor of 73.2%. So far, it is unclear whether FFM equations based on adults could be adapted for use in children by correction factor for hydration.
DXA is still the gold standard to measure body composition (29). The current evaluation, in line with others (25,30), neglects measurement error by DXA methodology (31) and ascribes all error, being the deviation between BIA and DXA, to the BIA methodology. Measurement error in FM (%) can also result from inaccurate detection of FM in the trunk region or variation in tissue thickness (28). Alternatively, an individual subject may have an overestimation of FFM by DXA and at the same time an underestimation by BIA-FFM (32). Strengths of this study include the use of DXA, a robust and well-accepted measure (29)and the systematic literature search of FFM-BIA equations.
A limitation of the study is that the interpretation is not applicable to other BIA and DXA devices or software. Besides this our DXA had a limit of 125 kg. Larger
118