Evaluation of Skinfolds Based Predictive Equations for Estimating Body Fat Percentage in Older Adults: A Comparison Using Digital Adipometer (Lipowise) and Dual-Energy X-Ray Absorptiometry (DXA).

Accurate assessment of body fat percentage (%BF) is essential for predicting health risks in older adults, as age-related changes affect fat distribution and measurement accuracy. While dual-energy x-ray absorptiometry (DXA) is a standard reference, it is costly and less accessible. This pilot study evaluated the validity of four skinfold equations using a digital calliper, with DXA as the reference to estimate %BF in older adults. A cross-sectional study included 149 participants (28 men and 121 women) aged 60 to 86 years. The Lean, Durnin-Womersley, and Visser equations (using 2 and 4 skinfolds) were analysed to estimate %BF. Statistical analyses were performed using SPSS (v.29), including Student's t-test, Mann-Whitney, Cohen's Kappa, and Bland-Altman. Skinfold equations differed significantly from DXA-derived %BF. In women, DXA values were lower than Lean but higher than Durnin-Womersley; in men, DXA showed lower %BF than Lean and higher than others. Agreement with DXA was moderate (Kappa 0.54-0.59). Durnin-Womersley underestimated %BF, especially in men (11.5%); Lean slightly overestimated (1.3%). Visser showed minimal bias-slight overestimation in women (0.3%, 0.4%) and underestimation in men (3.7%, 4.3%). Age-related physiological factors may impact equation accuracy. Discrepancies between DXA and skinfold equations highlight the need for age-specific prediction models for older adults.