Longitudinal measurements of total body water and body composition in healthy volunteers by online breath deuterium measurement and other near-subject methods.

Rapid quantification of breath deuterium abundance by flowing afterglow mass spectrometry (FA-MS) enables accurate measurement of total body water (TBW), which combined with other techniques such as bioelectrical impedance analysis (BIA) and anthropometrics enables near-subject assessment of body composition. This study assessed the comparative reproducibility and inter-relationship of these methods in healthy subjects over 12 months. Detailed bedside composition was performed in 22 subjects, (10 male) aged 28-79 with body mass index (BMI) ranging from 21-38 at baseline and again at one year. Techniques included FA-MS deuterium dilution, BIA, skin-fold thickness (SFT) and soft tissue ultrasound measurement of fat and muscle depth. Short-term reproducibility for each method was established. Within and between technique comparisons of measurement were made from Pearson's linear regression, coefficient of variation (CV) and Bland-Altman analysis. Weight and TBW estimated by FA-MS, BIA and SFT at baseline and one year later were highly correlated (R2 = 0.96-0.98), slope 1.02-1.03, CV = 4.5-11.6%. Systematic errors between the different methods in determining TBW were effectively identical at baseline and after one year. There was a tendency for subjects to gain weight during the study period, due to an increase, predominantly in younger women, of body water (FA-MS and SFT) and loss of upper body fat (ultrasound). BIA was relatively insensitive to these changes. It is concluded that over a 12-month period, TBW determined by FA-MS deuterium breath analysis has reproducibility similar to conventional weighing. The stability of between method errors would suggest that these techniques might be used in conjunction with each other in the longitudinal determination of body composition and so detect relatively subtle changes. The value of including an absolute determinant of TBW by FA-MS that is independent of the need to employ population derived equations, appears to be of value in the near-subject determination of body composition as required in clinical practice.