Accuracy of three novel predictive methods for measurements of fat mass in healthy older subjects

Abstract

This study evaluated the agreement of novel anthropometric equations and established indirect methods (skinfold thickness and bioimpedance analysis) with reference methods [dual X-ray absorptiometry (DXA) and air displacement plethysmography (ADP)] for fat mass assessment (FM) in older subjects. Forty subjects (M/F = 15/25, age = 61-84 years, BMI = 18-37 kg/m(2)) were recruited. The agreement of the following predictive equations was evaluated: body adiposity index (BAI), BAI-Fels and Clínica Universidad de Navarra-body adiposity estimator (CUN-BAE). BAI estimates were comparable to DXA (Δ ± 2SD = 0.4 ± 6.0 kg, p > 0.05) but not to ADP (Δ ± 2SD = -2.8 ± 7.2 kg, p < 0.001); BAI-Fels estimates were comparable to DXA (Δ ± 2SD = 0.8 ± 5.5 kg, p > 0.05) but not to ADP (Δ ± 2SD = -4.0 ± 6.9 kg, p < 0.001). The difference between CUN-BAE and ADP was not significant (Δ ± 2SD = -0.4 ± 5.6 kg, p > 0.05), whereas it significantly overestimated DXA (Δ ± 2SD = 2.8 ± 5.4 kg, p < 0.001). ADP significantly overestimated FM compared to DXA (Δ ± 2SD = 3.2 ± 5.4 kg, p < 0.001) and the measurement bias was significantly correlated with BMI in men (p = 0.004). The accuracy of the three anthropometric indexes is dependent on the choice of the reference method. The variability of the FM estimates was large and these indexes cannot be recommended for the assessment of FM in older subjects.