Accuracy Of Multi-frequency BIA In %Fat Change During Weight Loss Among Competitive Girl Runners

Abstract

Periodical measurement of %fat with accuracy is crucial for optimizing the health and performance in competitive runners who need strict weight control. Multi-frequency bioelectrical impedance analysis (MFBIA) is feasible for routine use but the evidence of the accuracy is limited to cross-sectional evaluation and no data is available for tracking of %fat change in competitive runners. PURPOSE: To determine the accuracy of MFBIA for evaluating the decrease in %fat by weight loss among competitive girl runners. METHODS: The data of %fat were obtained from consecutive 25 freshmen long distance runners over 5 years in the same girlʼs high school team and were retrospectively analyzed. The team regularly participated in the All-Japan high school Ekiden championship. MFBIA was performed at the preparatory season (PRE) and repeated after 5.6 ± 0.5 months at the competitive season (CMP) with dual energy X-ray absorptiometry (DXA) as reference. Weight loss period was defined as that between PRE and CMP and changes in %fat (Δ%fat) was calculated by subtracting %fat in PRE from those in CMP. Bland-Altman analysis was used to evaluate the validity of MFBIA compared to DXA for Δ%fat. Statistical significance of the mean difference between MFBIA and DXA was assessed by paired t-test. P<0.05 was considered as statistically significant. Written informed consent was obtained from the runners and their parents. RESULTS: %Fat by DXA vs. MFBIA at PRE and CMP were 19.7 ± 5.6 vs. 17.9 ± 5.4% and 12.7 ± 3.1 vs. 12.2 ± 2.6%, respectively. Thus, systemic error (the mean difference) of Δ%fat between the 2 methods was 1.2 percentage points [pp] (DXA, -6.9 vs. MFBIA, -5.7pp, p=0.008). Random error (the limits of agreement) was -3.0 to 5.5pp and no proportional error was observed between MFBIA and DXA. CONCLUSIONS: The small size of systemic error (mean difference) allow the use of MBIA to evaluate a group mean of Δ%fat. However, due to the large random error size relative to the low level of %fat of the competitive runners, caution should be taken to use MFBIA for individual monitoring of %fat change during weight loss period.