Age and sex differences in the modification of the leg lean mass-bone mineral density association by sedentary time in U.S. adults

Background Lean mass is positively associated with bone mineral density (BMD), and this relationship may vary according to age and sex. Sedentary behavior has also been associated with BMD, although findings differ across populations and skeletal sites. However, it remains unclear whether sedentary time modifies the relationship between regional lean mass and BMD and whether such modification varies across sex and age. These relationships were explored using nationally representative U.S. data. Methods Adults aged 18-59 years from the 2017-2018 National Health and Nutrition Examination Survey (NHANES) were analyzed. Mean bilateral leg lean mass and mean bilateral leg BMD were derived from dual-energy X-ray absorptiometry measurements, and sedentary time was obtained from self-report. Survey-weighted regression models incorporated NHANES examination weights, strata, and primary sampling units. The primary model examined the four-way interaction among leg lean mass, sedentary time, sex, and continuous age, with adjustment for race/ethnicity and height. Quadratic lean-mass terms were examined in sensitivity analyses. Results The primary complete-case analysis included 2,543 adults (1,191 men and 1,352 women). Women showed a more apparent decline in both leg lean mass and leg BMD at ages 50-59 years, whereas sedentary time did not increase in parallel with these age-related changes. Leg lean mass was positively associated with leg BMD in both sexes, with modest nonlinearity, particularly among women. In the primary survey-weighted model, the four-way interaction among leg lean mass, sedentary time, sex, and age was positive (beta = 0.000111 per year, 95% CI 0.000039-0.000183; P = 0.0051), indicating increasing sex divergence in sedentary-time modification of the lean mass-BMD association with age. At younger ages, conditional interaction estimates were similar between sexes; at older ages, estimates became increasingly positive in women while remaining negative but imprecise in men. Quadratic sensitivity analyses attenuated the four-way interaction estimate by approximately 9-12% and reduced its statistical precision, while broadly preserving its direction and magnitude. Conclusions In this exploratory cross-sectional analysis, sedentary-time modification of the association between DXA-derived leg lean mass and leg BMD varied according to sex and age. The sex difference was small at younger ages and became progressively more apparent with increasing age. These findings suggest that behavioral and demographic context may contribute to variation in the muscle-bone relationship itself, although confirmation in independent and longitudinal datasets is needed.