Adiposity and Muscle Mass Phenotyping is not Superior to BMI in Detecting Cardiometabolic Risk in a Cross-Sectional Study

Objective Classifying adiposity based on dual-energy x-ray absorptiometry (DXA) muscle and fat mass phenotypes has been proposed. Whether these phenotypes are more accurate in predicting cardiometabolic risk than BMI weight status is unknown. Methods Data were from the National Health and Nutrition Examination Survey (NHANES; 1999-2006 cycles, n = 5,475). Weight status was defined by BMI. Phenotypes of adiposity and muscle were based on high (>= 50th percentile) and low (<50th percentile) permutations of sex- and age-specific fat and muscle mass population curves. The area under the curves of receiver operating characteristic curves (ROC-AUCs), which predicted the presence of abnormal lipids, glucose, and blood pressure, were compared. All analyses were stratified by sex and incorporated the complex survey design and weighting of NHANES. Results The ROC-AUCs from weight status models used to correctly identify cardiometabolic risk ranged from 0.57 to 0.68, indicating generally weak predictive power. However, the ROC-AUCs from DXA phenotypes were lower (ranging from 0.53-0.68), indicating weaker predictive power than weight status, and were statistically inferior for nearly all of the comparisons. Conclusions Despite DXA's high cost and detailed output regarding body composition, its phenotype classification was inferior to weight status in predicting cardiometabolic risk. Further studies investigating the utility of the phenotypes are needed.

Automated summary

Analysis of NHANES data revealed that classifying adiposity based on DXA muscle and fat mass phenotypes was less accurate in predicting cardiometabolic risk compared to BMI weight status.