Estimation of Skeletal Muscle Mass and Visceral Adipose Tissue Volume by a Single Magnetic Resonance Imaging Slice in Healthy Elderly Adults

Abstract

Assessing skeletal muscle (SM) and visceral adipose tissue (VAT) by a single MRI slice at lumbar vertebra (L) 3 can replace whole-body MRI in young and middle-aged adults. However, this technique has not been proven in older adults. The aim of this analysis was to reinvestigate the best estimate for SM and VAT in an independent population of healthy elderly people. SM and VAT were assessed by whole-body MRI in 84 subjects ≥60 y [45 men; mean ± SD age: 68.4 ± 5.4 y, mean ± SD body mass index (in kg/m2): 25.5 ± 3.5]. SM and VAT areas of 9 slices at the lumbar spine were analyzed. The best estimate was investigated by Pearson correlations. Total volumes (in liters) were predicted by the area at lumbar vertebra 3 (AL3). Besides Bland-Altman analysis, linear regressions were performed to explain the variance of the bias by age, height, and percentage of fat mass (%FM). In a mixed population (healthy elderly plus reference population), linear regression with total SM and VAT volume as dependent variables and AL3, age, and height as independent variables was applied. When comparing the correlation coefficients between the tissue areas and total volumes, L3 was identified as the best estimate (r range: 0.71-0.94; all P < 0.05). However, Bland-Altman analysis showed a positive SM bias in men (mean ± SD: -1.0% ± 9.0%; P < 0.05) and a negative SM bias in women (mean ± SD: 3.7% ± 9.6%; P < 0.05). Contrary to SM, no significant bias was observed for VAT. In the elderly, stepwise linear regression showed height as a predictor for SM bias (R2 = 0.21, SEE = 2.07 L; P < 0.05) and %FM and age as predictors of the nonsignificant VAT bias (R2 = 0.26, SEE = 0.22L, P < 0.05), in men only. In the mixed population, AL3 and height were predictors for total SM, and AL3 for total VAT, independent of sex. AL3 was confirmed as the best estimate for SM and VAT volumes in healthy elderly adults. Contrary to VAT, there is a bias for SM, and height has to be added to the algorithm.