Department of Radiology, University Hospital, LMU Munich, Germany.
Institute of Clinical Radiology and Nuclear Medicine, University Medical Center Mannheim, Mannheim, Germany.
Br J Radiol. 2020 Jun;93(1110):20190874. doi: 10.1259/bjr.20190874. Epub 2020 Mar 27.
MRI is established for measurement of body fat mass (FM) and abdominal visceral adipose tissue (VAT). Anthropometric measurements and bioelectrical impedance analysis (BIA) have been proposed as surrogates to estimation by MRI. Aim of this work is to assess the predictive value of these methods for FM and VAT measured by MRI.
Patients were selected from cohort study PPS-Diab (prediction, prevention and subclassification of Type 2 diabetes). Total FM and VAT were quantified by MRI and BIA together with clinical variables like age, waist and hip circumference and height. Least-angle regressions were utilized to select anthropometric and BIA parameters for their use in multivariable linear regression models to predict total FM and VAT. Bland-Altman plots, Pearson correlation coefficients, Wilcoxon signed-rank tests and univariate linear regression models were applied.
116 females with 35 ± 3 years and a body mass index of 25.1 ± 5.3 kg/m were included into the analysis. A multivariable model revealed weight (β = 0.516, < 0.001), height (β = -0.223, < 0.001) and hip circumference (β = 0.156, = 0.003) as significantly associated with total FM measured by MRI. A additional multivariable model also showed a significant predictive value of FM (β = 0.583, < 0.001) for FM. In addition, waist circumference (β = 0.054, < 0.001), weight (β = 0.016, = 0.031) in one model and FM (β = 0.026, = 0.018) in another model were significantly associated with VAT quantified by MRI. However, deviations reached more than 5 kg for total FM and more than 1 kg for VAT.
Anthropometric measurements and BIA show significant association with total FM and VAT.
As these measurements show significant deviations from the absolute measured values determined by MRI, MRI should be considered the gold-standard for quantification.
磁共振成像(MRI)被确立用于测量体脂肪量(FM)和腹部内脏脂肪组织(VAT)。人体测量学测量和生物电阻抗分析(BIA)已被提出作为 MRI 估计的替代方法。本研究旨在评估这些方法对 MRI 测量的 FM 和 VAT 的预测价值。
从队列研究 PPS-Diab(2 型糖尿病的预测、预防和分类)中选择患者。通过 MRI 与 BIA 以及临床变量(如年龄、腰围和臀围以及身高)一起定量测量总 FM 和 VAT。利用最小角回归选择人体测量学和 BIA 参数,用于多元线性回归模型,以预测总 FM 和 VAT。应用 Bland-Altman 图、Pearson 相关系数、Wilcoxon 符号秩检验和单变量线性回归模型。
纳入了 116 名女性,年龄 35 ± 3 岁,体重指数为 25.1 ± 5.3kg/m2。多元模型显示体重(β=0.516,<0.001)、身高(β=-0.223,<0.001)和臀围(β=0.156,=0.003)与 MRI 测量的总 FM 显著相关。另外一个多元模型还显示 FM 对 FM 有显著的预测价值(β=0.583,<0.001)。此外,一个模型中腰围(β=0.054,<0.001)、体重(β=0.016,=0.031),另一个模型中 FM(β=0.026,=0.018)与 MRI 定量的 VAT 显著相关。然而,这些偏差超过了总 FM 的 5kg,超过了 VAT 的 1kg。
人体测量学测量和 BIA 与总 FM 和 VAT 有显著的关联。
由于这些测量值与 MRI 确定的绝对测量值存在显著偏差,因此 MRI 应被视为定量的金标准。