Abstract
Background: The four-compartment (4C) model is a criterion method for evaluating body composition tools like bioelectrical impedance analysis (BIA). This systematic review assessed the clinical equivalence of BIA devices compared to the 4C model and explored limitations in using the 4C model as a criterion method. Methods: Twelve cross-sectional and baseline longitudinal studies involving healthy, weight-stable, non-athlete, non-pregnant adults were included. The primary outcome was a Bland-Altman analysis, with bias, limits of agreement, and proportional bias extracted from each paper. The study quality was evaluated using the AXIS tool. Due to the high variability across studies, a meta-analysis was not performed. Results: BIA devices generally performed poorly against the 4C model estimates of percentage body fat and fat-free mass. Across the 12 studies, mean bias for percentage body fat between BIA and the 4C model ranged from -3.5% to +4.4%, with limits of agreement typically spanning 15 to 20 percentage points. For fat-free mass, mean bias ranged from -3.9 kg to +1.8 kg, with limits of agreement often exceeding ±6 kg. These wide limits indicate non-equivalence at the individual level despite small mean differences. Differences in both BIA device design and variations in 4C methodology across studies may have contributed to these discrepancies. Conclusions: BIA estimates of percentage body fat and fat-free mass were overall not equivalent to the 4C model. Alternative criterion methods, such as MRI, and use of raw BIA data are recommended. Standardization of BIA devices is also needed for improved clinical and research use.