Agreement Between a Photograph-Based Five-Compartment Body Composition Model and a Three-Compartment Reference Among Trained Adults

Abstract Lafontant, K, Smith, S, Livingston, J, Gonzalez, C, Nguyen, NLN, Kampiyil, S, Da Silva Barbera, MA, Johnson, B, Stout, JR, and Fukuda, DH. Agreement between a photograph-based 5-compartment body composition model and a 3-compartment reference among trained adults. J Strength Cond Res XX(X): 000–000, 2026—Multi-compartment models are widely accepted as the current standard for body composition assessments, yet the need for laboratory-based tests can make these models inaccessible for practitioners. A multi-compartment model based on photogrammetry and bioelectrical impedance (Bennett 5C ) has recently been proposed, yet it has not been validated in another cohort. We assessed agreement between a Siri 3-compartment (Siri 3C ) reference and Bennett 5C with and without corrections for total body water (TBW; using bioelectrical impedance) and total body volume (TBV; using photogrammetry) in a sample of 52 adult high intensity functional training practitioners (men: n = 34, women: n = 18, age: 34.9 ± 11.2 years, BMI = 26.0 ± 3.5 kg/m 2 ). Bland-Altman analyses and concordance correlation coefficients (CCC) were used, with the Bonferroni-corrected alpha set at p < 0.017. With corrections, Bennett 5C tended to underestimate fat mass (bias = 1.48 kg), overestimate fat-free mass (bias = −1.48 kg), and underestimate body fat percentage (bias = 1.63%) at a statistically significant level ( p ≤ 0.014; CCC, 0.743–0.966) with wide limits of agreement. Without corrections, Bennett 5C significantly differed from Siri 3C ( p < 0.001; CCC, 0.395–0.856), and produced negative body composition values for 4 subjects. Corrected TBV had adequate agreement with the criterion measure of air displacement plethysmography (bias = 0.05 L; limits of agreement = −2.37, 2.48 L; CCC = 0.997; p = 0.763). When TBV and TBW estimates are corrected, practitioners can use Bennett 5C to estimate body composition within trained populations but should be wary of the wide limits of agreement and consider alternative equations.