The Influence of Subject Leanness on Statistical Agreement of Fit3D Parameters During Serial Scans

Abstract

Body composition influences numerous health outcomes. The Fit3D Body Scanner (Fit3D Inc., San Mateo, CA, USA) is camera-based system that estimates a variety of anthropometric parameters. Reports on the within-subject variance of its outputs are limited.

PURPOSE: To evaluate the agreement of Fit3D parameters on a bodybuilder during a period of consistent weight loss.

METHODS: We tested a 30-year-old female IFBB professional bodybuilder on the Fit3D system every Saturday for 10 consecutive weeks during contest preparation. Each session occurred at 6:30 AM following a 12-hour fast of food and fluids, and consisted of 6 sequential scans. We exported weight, waist circumference (WC), hip circumference (HC), lean mass, fat mass, body fat percent (BF%), trunk to leg volume ratio (TLVR), a body shape index (ABSI), surface-based body shape index (SBSI), and body shape rating (BSR). For each variable, we calculated mean, variance, standard deviation, and intra-class correlation coefficient (ICC; two-way mixed, absolute agreement).

RESULTS: Upon initial testing, weight was 107.1 ± 0.7 lb, WC was 28.5 ± 0.3 in, HC was 34.9 ± 0.2 in, lean mass was 87.0 ± 1.9 lb, fat mass was 20.1 ± 1.8 lb, BF% was 18.8 ± 1.7%, TLVR was 1.5 ± 0.1, ABSI was 0.1 ± 0.0, SBSI was 0.1 ± 0.0, and BSR was 62.3 ± 4.7. At week 10, weight was 96.6 ± 0.1 lb, WC was 25.4 ± 0.3 in, HC was 32.2 ± 0.2 in, lean mass was 86.8 ± 5.5 lb, fat mass was 9.8 ± 5.5 lb, and BF% was 10.2 ± 5.7%, TLVR was 1.6 ± 0.4, ABSI was 0.1 ± 0.0, SBSI was 0.1 ± 0.0, and BSR was 70.4 ± 2.4. Reproducibility of these measurements ranged from excellent to poor: weight (ICC = 0.997; 95% CI: 0.994, 0.999), WC (ICC = 0.971; 95% CI: 0.932, 0.991), HC (ICC = 0.971; 95% CI: 0.934, 0.992), ABSI (ICC = 0.728; 95% CI: 0.494, 0.915), fat mass (ICC = 0.695; 95% CI: 0.450, 0.902), BF% (ICC = 0.572; 95% CI: 0.306, 0.848), BSR (ICC = 0.537; 95% CI: 0.268, 0.830), lean mass (ICC = 0.326; 95% CI: 0.089, 0.698), TLVR (ICC = 0.083; 95% CI: -0.073, 0.456), and SBSI (ICC = 0.072; 95% CI: -0.079, 0.440).

CONCLUSIONS: Satisfactory reproducibility was found in most Fit3D outputs evaluated. At a leaner composition, variances were larger in several features of body composition, including lean mass, fat mass, and BF%.