A population-based approach to define body-composition phenotypes123

Abnormal body compositions such as high adiposity (HA), low muscle mass (LM), or a combination of the 2 [high adiposity with low muscle mass (HA-LM)] are relevant phenotypes, but data on their prevalence and impact on health are still limited. This is largely because of a lack of a consensus definition for these conditions. Of particular interest is the HA-LM phenotype, also termed “sarcopenic obesity,” which may confer greater health risk. We propose a new approach for operationalizing abnormal body-composition phenotypes in a representative adult population. Whole-body dual-energy X-ray absorptiometry data obtained from the 1999–2004 NHANES were analyzed for 13,236 subjects aged ≥18 y (maximum weight and height of 136 kg and 1.96 m, respectively). Sex- and body mass index (BMI)–specific decile groups of appendicular skeletal muscle index (ASMI; kg/m2) and fat mass index (FMI; kg/m2) were developed. Cutoffs for HA and LM were incorporated into a diagnostic framework to characterize 4 specific body-composition phenotypes—low adiposity with high muscle mass, high adiposity with high muscle mass, low adiposity with low muscle mass, and HA-LM—and a subclassification of the phenotypes into classes I, II, and III. Abnormal phenotypes were prevalent across the age spectrum and BMI categories. The association between ASMI or FMI and age was modified by sex and BMI. The prevalence of HA-LM in the whole sample was 10.3% in women and 15.2% in men. The prevalence of all subclasses of HA-LM in obese women and men was 14.7% and 22.9%, respectively. HA-LM class III was more prevalent in obese men (2.3%) than in obese women (0.3%). We developed sex- and BMI-specific reference curves to harmonize the classification of body-composition phenotypes. The application of this classification will be particularly useful in the identification of cases of sarcopenic obesity. The association of these phenotypes with metabolic deregulation and increased disease risk awaits verification.