Abstract
OBJECTIVES: The impact of adiposity on mortality in older adults remains controversial. Some reports suggest that measures of general adiposity such as body mass index (BMI) predict better survival. We assessed the relationship between measures of adiposity and mortality in older adults. DESIGN: Cross-sectional analysis of a population-based sample. SETTING: Non-institutionalized persons in the United States participating in the National Health and Nutrition Examination Surveys III and its linked mortality dataset. PARTICIPANTS: A subsample of 4,489 non-institutionalized survey participants aged >60 years with measures of body composition using bioimpedance. To account for possible residual confounding, smokers, subjects with heart failure, respiratory disease, kidney disease and cancer were excluded (n=2,920). Data from 1569 subjects were analysed. MEASUREMENTS: BMI, waist circumference (WC), waist-hip ratio (WHR), lean mass (LM) and % Body Fat (BF) were classified by tertiles (lowest=referent). Proportional-hazard models evaluated the association of anthropometric indices with overall and cardiovascular mortality. RESULTS: Mean age was 69.4years, and 265(16.9%) were >80 years. There were 717(47.6%) women and 792 deaths of which 284 [35.9%] were cardiovascular related. Elevated BMI was associated with reduced cardiovascular mortality (HR 0.53 [0.30-0.84]), and remained significant after adjusting for LM (HR 0.54 [0.31-0.93]). Elevated %BF was associated with reduced mortality from cardiovascular causes (HR 0.52 [0.29-0.91]). Low BMI was associated with higher risk of cardiovascular (HR 3.66 [1.25-10.69]) and overall death (HR 2.44 [1.22-4.90]). CONCLUSION: Measures of adiposity in older participants are associated with lower mortality from cardiovascular causes that cannot be explained by major known confounders between obesity and mortality. Further studies need to elucidate a possible protective role and interplay between adiposity and skeletal muscle in older adults.