Sex differences in physical activity dose-response effects on site-specific bone mineral density during childhood and adolescence

Abstract

Physical activity (PA) serves as a modifiable determinant of bone mineral density (BMD). However, sex- and site-specific dose-response relationships remain poorly defined. Leveraging data from the National Health and Nutrition Examination Survey (NHANES 2011-2014), this cross-sectional study investigated sex-specific associations between accelerometer-derived PA (volume and intensity) and site-specific BMD in 2,659 children and adolescents aged 8-19 years. BMD assessment was conducted via dual-energy X-ray absorptiometry, and PA was quantified using monitor-independent movement summary units (MIMS). Linear and nonlinear analyses revealed distinct patterns: boys exhibited stronger linear associations between PA volume and BMD at weight-bearing sites (e.g., pelvis, arms, and legs; β = 0.003-0.004, P < 0.05), and girls demonstrated nonlinear thresholds (e.g., volume threshold: 15.0 × 10³ MIMS/day). Both sexes had intensity-driven thresholds (~ 45-49 MIMS/min), with diminishing returns above these values (β = 0.010 in boys; stagnation in girls). Directional trends suggest the greater BMD gains per standard deviation (SD) increase in PA for boys compared with that for girls across most skeletal regions (e.g., total body less head: 0.131 SD vs. 0.106 SD for intensity). However, statistical significance (P < 0.05) was observed only for arm BMD responses. Mechanistic analyses highlighted the corresponding biomechanical principles, with weight-bearing regions showing stronger PA-BMD links than nonweight-bearing sites (e.g., spine). These findings underscore the importance of prioritizing high-intensity, weight-bearing activities, particularly for girls, to optimize skeletal health during growth. Public health strategies should focus on achieving, rather than exceeding, the identified daily 60 min intensity thresholds (~ 45-49 MIMS/min) through school- and community-based interventions.

Keywords: Bone mineral density; Dose-response; Pediatric; Physical activity; Sex differences..

Fig. 1

Participant selection flowchart. ^*Covariates analyzed: age, sex, race/ethnicity, height, weight, poverty income ratio, dietary calcium, and vitamin D. PA, physical activity; BMD, bone mineral density; NHANES, National Health and Nutrition Examination Survey.

Fig. 2

Age-related BMD distribution across skeletal sites (8–19 years). Weighted mean BMD values with standard errors are represented by data points. ‌Significant sex differences (Benjamini-Hochberg adjusted P < 0.05) are annotated with red circles along the x-axis. Regions: (a) TBLH, (b) thoracic spine, (c) lumbar spine, (d) pelvis, (e) arms, (f) legs. BMD, bone mineral density; TBLH, total body less head.

Fig. 3

Nonlinear threshold effects of PA volume and intensity on BMD at select skeletal sites. Panels display regions with statistically significant nonlinear associations (P < 0.05): ‌Volume effects (per 10³ MIMS/day):‌ (a) TBLH, (b) pelvis, (c) legs, (d) lumbar spine; ‌Intensity effects (per 10 MIMS/min):‌ (e) TBLH, (f) pelvis, (g) legs, (h) arms. PA, physical activity; BMD, bone mineral density; MIMS, monitor-independent movement summary units; TBLH, total body less head.

Fig. 4

Sex- and site-specific associations between PA and BMD. Associations are presented as standardized β-values (95% CIs) based on weighted linear regression comparing z-score normalized PA and BMD measurements. PA, physical activity; BMD, bone mineral density; SD, standard deviation; TBLH, total body less head.