Objective: To investigate whether within-individual variation is a factor in the generally reported poor tracking of central body fat distribution (CBFD) during development and whether two measurements of CBFD during each measurement occasion would improve the estimate of tracking over time.
Methods: A longitudinal study compared the results of two measurements of body fat (BF) and CBFD during each measurement occasion to the results of one measurement of BF and CBFD during each occasion every 4 months over 1-3 y. A total of 345 boys and 333 girls in three age cohorts of 8, 11 and 14 y at baseline were examined. BF variables were: body mass index (BMI); fat mass and percentage body fat from bioelectrical impedance; two skinfold sums; and abdominal circumference. CBFD variables were: waist/hip and waist/thigh ratios; conicity; and log upper-lower skinfold ratio.
Results: Three-year tracking of BF varied from 0.79 to 0.90 for one- and from 0.81 to 0.93 for two-occasion measurements showing no apparent sex- or cohort-effects. Three-year tracking of CBFD was lower than that of BF (0.68-0.75), but improved significantly with two-occasion measurements (0.75-0.82).
Conclusions: Within-individual variation is a significant factor in reported low tracking of CBFD in childhood. Estimates of tracking currently in the literature may underestimate the predictive value of CBFD, possibly because this research has used one-occasion measurement. The increased use of two-occasion measurement should significantly improve the tracking of CBFD during development and provide a more realistic understanding of its predictive value.