Learning Effects in Air Displacement Plethysmography

Air displacement plethysmography (ADP) is a widespread technique for assessing global obesity in both health and disease. The reliability of ADP has been demonstrated by studies focused on duplicate trials. The present study was purported to evaluate learning effects on the reliability of body composition assessment using the BOD POD system, the sole commercially available ADP instrument. To this end, quadruplicate trials were performed on a group of 105 subjects (51 women and 54 men). We estimated measurement error from pairs of consecutive trials-(1,2), (2,3), and (3,4)-to test the hypothesis that early measurements are subject to larger errors. Indeed, statistical analysis revealed that measures of reliability inferred from the first two trials were inferior to those computed for the other pairs of contiguous trials: for percent body fat (%BF), the standard error of measurement (SEM) was 1.04% for pair (1,2), 0.71% for pair (2,3), and 0.66% for pair (3,4); the two-way random effects model intraclass correlation coefficient (ICC) was 0.991 for pair (1,2), and 0.996 for pairs (2,3) and (3,4). Our findings suggest that, at least for novice subjects, the first ADP test should be regarded as a practice trial. When the remaining trials were pooled together, the reliability indices of single ADP tests were the following: ICC = 0.996, SEM = 0.70%, and minimum detectable change (MDC) = 1.93% for %BF, and ICC = 0.999, SEM = 0.49 kg, and MDC = 1.35 kg for fat-free mass (FFM). Thus, the present study pleads for eliminating learning effects to further increase the reliability of ADP.

The impact of subject positioning on body composition assessments by air displacement plethysmography evaluated in a heterogeneous sample

Introduction

This study sought to evaluate the impact of subject positioning on body composition assessments by air displacement plethysmography using the BOD POD®.

Methods

Eighty-two adults (42 men and 40 women), aged 26.1 ± 8.4 y (mean ± standard deviation), body mass index = 23.6 ± 4.8 kg/m2, were assessed by repeated measurements in two different positions: relaxed (legs apart, back away from the rear) and compact (legs together, arms near the body, back touching the rear). We relied on Bland-Altman analysis to quantify the agreement between results recorded in the two positions. Using body surface charts, we tested the hypothesis that posture-induced variability stems from differences in exposed skin area.

Results

Switching from compact to relaxed position resulted in a bias of -197 mL for body volume, -1.53% for percent body fat, and 1.085 kg for fat-free mass. The body surface area in contact with air was larger in relaxed position by 3632 ± 522 cm2. When body volume was expressed in terms of the actual area of exposed skin in the compact position, the percent body fat bias became 0.08%, with a 95% confidence interval of (-0.14, 0.29)%.

Conclusions

Subject posture is a source of significant variability in air displacement plethysmography. The disagreement between results obtained in different positions can be eliminated by adjusting the surface area artifact, suggesting that subject positioning in the BOD POD® should be controlled to avoid changes in the amount of air maintained under isothermal conditions by the body.