Predictive model specific to young adults for estimating thoracic gas volume for air-displacement plethysmography

Background Thoracic gas volume either measured (mTGV) or predicted by the BodPod (R) (bpTGV) is used during air-displacement plethysmography to obtain a better estimate of percent body fat. Evidence suggests that bpTGV underestimates mTGV for young adults and this is especially evident for young males. Aims We developed, validated, and cross-validated a TGV prediction model (pTGV) for males and females 18-30 years of age to address this underestimation. Materials & Methods Participants (N = 181; 18-30 years) that had their body composition assessed with the BodPod (R) were retrospectively randomly assigned to one of two independent subgroups, a validation (n = 145) or cross-validation (n = 36) sample. Ten iterations of the k-fold validation procedure were performed to assess the internal replicability of pTGV within the validation sample. External replicability of pTGV was evaluated by assessing the difference and standard error of the estimate (SEE) compared to mTGV in the cross-validation group. Results The model using height, sex and body mass yielded the highest adjusted R-2 (0.627) and the lowest SEE (0.56 L): pTGV = 0.615338 x Sex (0 = Female, 1 = Male) + 0.056267 x Height (cm) - 0.011006 x Body Mass (kg) - 5.358839. R-2 remained stable across 10 iterations of the k-fold procedure (average R-2 = 0.64). Differences between pTGV and mTGV were not significantly different than zero for the total cross-validation sample (-0.06 +/- 0.7 L; SEE = 3.0%), for males (-0.11 +/- 0.7 L; SEE = 3.7%), or for females (-0.02 +/- 0.7 L; SEE = 5.3%). Conclusion We recommend that when it is impractical to obtain mTGV, the strong internal and external replicability of the new prediction model supports its use for males and females ages 18-30 years old during air-displacement plethysmography.