Two-megahertz impedance index prediction equation for appendicular lean mass in Korean older people

Abstract

Background Whole-body bioelectrical impedance analysis (BIA) has been accepted as an indirect method to estimate appendicular lean mass (ALM) comparable to dual-energy X-ray absorptiometry (DXA). However, single or limited frequencies currently used for these estimates may over or under-estimate ALM. Accordingly, there is a need to measure the impedance parameter with appendicular lean-specific across multiple frequencies to more accurately estimate ALM. We aimed to validate muscle-specific frequency BIA equation for ALM using multifrequency BIA (MF-BIA) with DXA as the reference. Methods 195 community-dwelling Korean older people (94 men and 101 women) aged 70 ~ 92y participated in this study. ALM was measured by DXA and bioimpedance measures at frequencies of 5 kHz ~ 3 MHz were assessed for independent predictive variables. Regression analyses were used to find limb-specific frequencies of bioimpedance, to develop the ALM equations and to conduct the internal cross-validation. The six published equations and the final equation of MF-BIA were externally cross-validated. Results 195 participants completed the measurements of MF-BIA and DXA. Using bivariate regression analysis, the 2 MHz impedance index explained R2 = 91.5% of variability (P < 0.001) in ALM and predictive accuracy of standard error of estimate (SEE) was 1.0822 kg ALM (P < 0.001). Multiple stepwise regression analysis obtained in the development group had an adjusted R2 of 9.28% (P < 0.001) and a SEE of 0.97 kg ALM. The cross-validation group had no significant difference between the measured ALM and the predicted ALM (17.8 ± 3.9 kg vs. 17.7 ± 3.8 kg, P = .486) with 93.1% of R2 (P < 0.001) and 1.00 kg ALM of total error. The final regression equation was as follows: ALM = 0.247ZI@2 MHz + 1.254SEXM1F0 + 0.067Xc@5 kHz + 1.739 with 93% of R2 (P < 0.001), 0.97 kg ALM of SEE (Subjective Rating as “excellent” for men and “very good” for women). In the analysis of the diagnostic level for sarcopenia of the final regression, the overall agreement was 94.9% (k = 0.779, P < 0.001) with 71.4% of sensitivity, 98.8% of specificity, 91.3 of positive prediction value and 95.3% of negative prediction value. Conclusion The newly developed appendicular lean-specific high-frequency BIA prediction equation has a high predictive accuracy, sensitivity, specificity, and agreement for both individual and group measurements. Thus, the high-frequency BIA prediction equation is suitable not only for epidemiological studies, but also for the diagnosis of sarcopenia in clinical settings.

Body Composition Assessment of University Athletes: Comparison Between the Data Obtained by Bioelectrical Impedance and by Anthropometry

“…The interpretation of data evaluated by the BIA begins to be dubious due to the fact that there are studies that do not demonstrate correlations of results between different BIA devices (Kim et al 2022, Lee et al 2017. For example, the predictive equations of the InBody230® scale for evaluating FM and SMM are not publicly available and cannot be manipulated by users.…”

High-standard predictive equations for estimating body composition using bioelectrical impedance analysis: a systematic review

“…Risk of bias was high in two [34,35] and unclear in six [36][37][38][39][40][41] studies due to the low to moderate (R 2 < 80%) coefficient of determination (R 2 ) resulted from the multiple regression models. Concerns with applicability were high in 19 studies [37,[42][43][44][45][46][47][48][49][50][51][52][53][54][55][56][57][58][59] and unclear [41,50,[60][61][62][63][64][65][66][67][68][69] in 12 studies due to the restricted sample size.…”

“…A total of 885 males and 396 females were included in 11 studies [31,39,48,51,62,69,[83][84][85][86][87][88] aimed to develop predictive equations for athletes. A total of 1469 males and N = 2111 females were included in 17 studies [36,38,46,53,60,63,67,68,[89][90][91][92][93][94][95][96][97] aimed to develop predictive equations for elderly people. A total of 358 males, 637 females, and 75 subjects without sex designated were included in 8 studies [35,40,41,56,64,65,98,99] aimed to develop predictive equations for people with diseases.…”

Correction: Two-megahertz impedance index prediction equation for appendicular lean mass in Korean older people

“…After publication of this article [ 1 ], the authors reported that in this article the author Fang Lin Jiang was incorrectly linked to affiliations 6 and 7, but should have been linked to affiliation 6 only; and the author Chul-Hyun Kim was incorrectly linked to affiliation 7, but should have been linked to affiliation 6.…”