Comparison of body composition techniques before and after a 161-km ultramarathon using DXA, BIS and BIA

The low cost, ease of application and portability of bioelectrical impedance analysis (BIA) and spectroscopy (BIS) devices make them attractive tools for measuring acute changes in body composition before and after exercise, despite potential limitations from active compartmental fluid shifts. The primary study aim was to evaluate use of dual energy x-ray absorptiometry (DXA) against BIA and BIS in measurements of percent body fat (%BF) and percent total body water (%TBW) before and after prolonged endurance exercise. 10 runners were measured pre-race and at race finish. Significant linear relationships were noted pre-race between DXA vs. BIS for %BF (r(2)=0.76; p<0.01) and %TBW (r(2)=0.74; p<0.01). Significant correlations were noted at race finish between DXA vs. BIS for %BF (r(2)=0.64; p<0.01) and %TBW (r(2)=0.66; p<0.05), but only when one outlier was removed. Limits of agreement (LOA) between DXA vs. BIS were wide for both %BF (mean difference of -3.6, LOA between 5.4 and -12.6) and %TBW (mean difference 2.4, LOA between 0.4 and -4.6). LOA was closer between the DXA vs. BIA with DXA measuring slightly higher than BIA for %BF (mean difference of 0.5, LOA between 2.1 and -3.1) and slightly lower than BIA for %TBW (mean difference 0.3, LOA between 3.3 and -2.7). Linear correlations between DXA vs. BIA were not statistically significant for %BF or %TBW before or after the race. DXA measurement of acute changes in %BF and %TBW are not congruent with BIA or BIS measurements. These 3 techniques should not be utilized interchangeably after prolonged endurance running.