Comparison of body composition measures assessed by bioelectrical impedance analysis versus dual-energy X-ray absorptiometry in the United Kingdom Biobank

Background: Bioelectrical impedance analysis (BIA) and dual-energy X-ray absorptiometry (DXA) serves as common modalities for body composition assessment. This study was aimed to evaluate the agreement between BIA and DXA measures in UK Biobank.

Methods: UK Biobank participants with body fat mass (FM) and fat-free mass (FFM) estimates obtained through BIA (Tanita BC418MA) and DXA concurrently were included. Correlation between BIA and DXA-derived estimates were assessed with Lin's concordance correlation coefficients. Bland-Altman and Passing-Boblok analyses were performed to quantify the difference and agreement between BIA and DXA. Multivariable linear regression was used to identify predictors influencing the differences. Finally, prediction models were developed to calibrate BIA measures against DXA.

Results: The analysis included 34437 participants (female 51.4%, mean age 64.1 years at imaging assessment). BIA and DXA measurements were highly correlated (Lin's concordance correlation coefficient 0.94 for FM and 0.94 for FFM). BIA (Tanita BC418MA) underestimates FM overall by 1.84 kg (23.77 vs. 25.61, p < 0.01), and overestimated FFM overall by 2.56 kg (52.49 vs. 49.93, p < 0.01). The BIA-DXA differences were associated with FM, FFM, BMI and waist circumference. The developed prediction models showed overall good performance in calibrating BIA data.

Conclusion: Our analysis exhibited strong correlation between BIA (Tanita BC418MA)- and DXA-derived body composition measures at a population level in UK Biobank. However, the BIA-DXA differences were observed at individual level and associated with individual anthropometric measures. Future studies may explore the use of prediction models to enhance the calibration of BIA measures for more accurate assessments in UK Biobank.