Longitudinal Assessment of Body Composition and Its Association With Survival Among Participants of the ACTIVE/ADIPOSE Study.

OBJECTIVES: The importance of muscle wasting as a predictor of mortality in the hemodialysis population is not clear. Lack of association of muscle mass with survival in some studies could be related to reliance on single measures or to incorporation of excess extracellular water (ECW) into estimates of muscle mass. We examined changes in body composition over a 2-year period and the association of body composition with survival. DESIGN AND METHODS: We analyzed data from 325 adults receiving hemodialysis in the Bay Area. We estimated ECW, intracellular water (ICW), and fat mass by whole-body bioimpedance spectroscopy (BIS) at 0, 12, and 24 months from enrollment. We used linear mixed modeling to examine changes in body mass index and BIS-derived estimates of body composition and Cox modeling with BIS-derived estimates as time-varying independent variables to examine associations between body composition and survival in multivariable analyses. RESULTS: Body mass index declined over time. Considering individual components of body composition, ICW declined (-0.09 kg/m2 per year, 95% confidence interval -0.14 to -0.04), but fat mass and ECW did not change significantly. There were 120 deaths over a median of 5.2 years. The relationship between ICW and mortality was not linear such that the association was steeper at low values of ICW, whereas higher ICW was associated with better survival that was relatively stable above 9 kg/m2. Higher ECW was associated with higher mortality, and fat mass was not associated with survival. These associations were independent of markers of inflammation and nutritional status. CONCLUSIONS: ICW declined over 2 years in this cohort, whereas fat mass and ECW remained relatively stable. Higher ICW was associated with better survival, but higher fat mass was not. Higher ECW was associated with worse survival. These results suggest that muscle mass may predict survival among patients on hemodialysis.