Objectives: To define the limits of change in body weight and body composit
ion after different time intervals in healthy, normal adults. Methods: Pros
pective and retrospective analyses of paired body composition studies in a
total of 326 healthy adults, ages 18 to 97. Measurements included body weig
ht, fat and fat-free mass (FFM) by dual x-ray absorptiometry (DXA) and bioi
mpedance analysis (BLA), plus body cell mass (BCM) by whole-body counting o
f K-40 and BIA. Results: Time interval between studies was a significant pr
edictor of the differences in paired studies. The 95% confidence intervals
for percent difference were lowest for body weight, intermediate for BCM an
d FFM, and highest for fat, in part because of the differences in sizes of
these body compartments. There were significant associations among the chan
ges in body composition by BIA and by criterion methods, suggesting that th
e observed changes are real. Conclusions: The normal variation in body weig
ht and body composition increases over time. Time-dependent criteria may in
crease the sensitivity in diagnosing malnutrition. Interpreting changes in
body compartments requires consideration of the size of each compartment.