Mediating effect of fat mass, lean mass, blood pressure and insulin resistance on the associations of accelerometer-based sedentary time and physical activity with arterial stiffness, carotid IMT and carotid elasticity in 1574 adolescents

Arterial stiffness, carotid intima-media thickness (cIMT) and carotid elasticity are surrogate markers of subclinical arteriosclerosis and atherosclerosis linked to cardiometabolic morbidity and mortality. While accelerometer-based sedentary time (ST), light physical activity (LPA) and moderate-to-vigorous physical activity (MVPA) have informed pediatric metabolic health guidelines, evidence for their relationships with arterial structure and function in youth is limited due to few studies using gold-standard vascular and objective activity measures. Body composition, BP, and metabolic markers are determinants of vascular health in youth, but whether they mediate associations between movement behaviors and arterial structure/function is unclear. This study investigated whether total body fat mass, trunk fat mass, lean mass, BP, and insulin resistance mediate the associations of ST, LPA and MVPA at age 15 years with cfPWV, cIMT and carotid elasticity at age 17 years in 1574 adolescents from the ALSPAC cohort.

Previous studies show mixed associations of sedentary time and physical activity with vascular measures in youth and young adults. Some adult/young adult cohorts report ST associated with higher cfPWV with mediation by adiposity and BP, whereas others found no association. Evidence on LPA and cfPWV in the young is sparse and previously null in smaller samples. MVPA effects on cfPWV have been inconsistent, including null findings in randomized trials among youth and young adults. Few studies have assessed cIMT or carotid elasticity in relation to objectively measured movement behaviors. Lean mass has emerged as an important determinant of vascular remodeling in youth, whereas BMI-based proxies may misattribute effects. This study addresses gaps by using objective accelerometry, gold-standard DEXA body composition, ultrasound/Vicorder vascular assessments, and formal mediation analyses in a large adolescent cohort.

Design and cohort: Prospective cohort analysis within ALSPAC (Avon Longitudinal Study of Parents and Children), UK. Participants: 1574 adolescents with valid accelerometer-derived ST, LPA, MVPA at age 15 years and vascular measures at age 17 years. Ethics approvals obtained; informed consent as per ALSPAC protocols. Physical activity and sedentary time: ActiGraph AM7164 2.2 accelerometer worn on the waist for 7 days at age 15. Valid day: ≥10 h wear time (excluding ≥10 min consecutive zero counts). Inclusion required ≥3 valid days (≥2 weekdays, ≥1 weekend day). Data processed with Kinesoft v3.3.75. Evenson cut-points used: ST 0–<100 counts/min; LPA 100–2296 cpm; MVPA >2296 cpm. ST and LPA categorized into sex-specific tertiles (low, middle, high). MVPA categorized as <40, 40–<60, and ≥60 min/day. Anthropometry and body composition: Height and weight at age 15; BMI calculated. DEXA assessed total body fat mass, trunk fat mass, and lean mass at age 15. Pubertal status was determined using time to age at peak height velocity via SITAR-based growth curve modeling. Vascular phenotypes (age 17): cfPWV measured using Vicorder by two trained technicians (inter-observer mean difference 0.2 m/s, SD 0.1). cIMT measured bilaterally on common carotids via ultrasound (12 MHz transducer; coefficient of variation 4.4 ± 2.2%); mean of right and left averages used. Carotid elasticity computed as systolic–diastolic diameter difference. Cardiometabolic and lifestyle factors (age 15): Resting heart rate; systolic and diastolic BP via Omron monitor; hypertensive systolic BP defined as ≥130 mmHg. Fasting blood: glucose, insulin, hs-CRP, LDL-C, HDL-C, triglycerides (CV <5%). HOMA-IR = fasting insulin × fasting glucose / 22.5. Smoking in last 30 days via questionnaire. Family history of hypertension, diabetes, high cholesterol, vascular disease at age 17. Socioeconomic status grouped per 1991 British OPCS. Statistical analysis: Descriptive statistics with appropriate tests for sex differences. Skewed variables log-transformed. Mediation analyses via structural equation models (1000 bootstrap samples) tested mediating roles of total fat mass, trunk fat mass, lean mass, systolic BP, diastolic BP, and HOMA-IR in associations of ST, LPA, MVPA (predictors) with cfPWV, cIMT, carotid elasticity (outcomes). Adjusted for age, sex, time between exposure and outcome, hs-CRP, heart rate, systolic BP, smoking, family history, socioeconomic status, HDL-C, LDL-C, triglycerides, lean mass or total fat mass as appropriate, glucose and insulin. Total, direct, and indirect effects estimated; mediation percentage computed. Linear mixed-effect models assessed categorical ST, LPA, MVPA associations with outcomes, with random intercepts and covariate adjustments as above, plus 20-cycle multiple imputation for missing covariates. Sex-stratified regressions and analyses stratified by systolic hypertension (≥130 mmHg) were performed. Collinearity checked (VIF <5). Two-sided p<0.05 significant with Sidak correction. Power considerations based on ALSPAC sampling.

Sample: 1574 adolescents; 56.2% female; mean age 15.4 years. MVPA ≥60 min/day: 40.1% males, 17.4% females. Sedentary time (ST): Higher ST associated with lower cIMT; lean mass partially mediated this association (11.8% mediation; Table 4). No significant mediation by total fat mass, trunk fat mass, systolic or diastolic BP, or HOMA-IR. Middle ST tertile associated with higher carotid elasticity (β=0.022; p=0.030) in mixed models; otherwise ST not associated with cfPWV. Light physical activity (LPA): Higher LPA associated with lower cfPWV (standardized β = -0.057; 95% CI -0.101 to -0.013; p=0.014) independent of mediators. Highest LPA tertile associated with lower cfPWV (β=-0.107; p=0.018). Moderate LPA tertile associated with higher carotid elasticity (β=0.026; p=0.010). No significant mediation by body composition, BP, or HOMA-IR. Sex-stratified: LPA associated with lower cfPWV in females (β=-0.048; p=0.044) but not males. In normotensive adolescents, LPA associated with lower cfPWV (β=-0.064; p=0.046) and higher carotid elasticity (β=0.078; p=0.018); no significant associations in hypertensive subgroup. Moderate-to-vigorous PA (MVPA): MVPA consistently associated with higher carotid elasticity across models (e.g., total effect standardized β≈0.073–0.089; p≤0.007). MVPA ≥60 min/day associated with higher carotid elasticity (β=0.024; p=0.027). MVPA was associated with higher cIMT when lean mass was included as mediator; lean mass mediated 25.5% of MVPA–cIMT association (Table 4). Lean mass mediated 28.1% of MVPA–carotid elasticity association. HOMA-IR mediated 7.7% of MVPA–carotid elasticity association. No significant mediation by total or trunk fat mass or by systolic/diastolic BP. Sex-stratified: MVPA associated with higher carotid elasticity in males (β=0.019; p=0.013), not in females. No significant associations of MVPA with cfPWV. Blood pressure as mediator: Systolic and diastolic BP did not significantly mediate relationships between movement behaviors and vascular indices. Overall: LPA was linked to better arterial function (lower cfPWV), whereas MVPA was linked to higher carotid elasticity and paradoxically higher cIMT via lean mass. ST showed a paradoxical association with lower cIMT, partially explained by lean mass.

The study addressed whether body composition, BP, and insulin resistance mediate relationships between accelerometer-measured movement behaviors and vascular structure/function in adolescence. Findings indicate distinct behavioral–vascular profiles: (1) higher ST was paradoxically associated with lower cIMT, with partial mediation by lean mass, suggesting that ST’s relationship to vascular structure may be influenced by muscle mass–related vascular remodeling; (2) higher LPA related to lower arterial stiffness (cfPWV) and greater carotid elasticity, particularly among normotensive youth and females, implying that LPA may beneficially impact arterial function without necessarily altering structure; (3) higher MVPA related to higher carotid elasticity and paradoxically higher cIMT when lean mass mediated the association, consistent with physiologic adaptations to increased muscle mass and repeated hemodynamic load. Total and trunk fat mass and BP did not mediate the observed associations, whereas lean mass and, to a lesser degree, insulin resistance, partially explained MVPA–elasticity links. These results refine understanding of how different intensities of movement affect adolescent vascular health and highlight lean mass as a key pathway. Public health implications include the potential value of promoting LPA for arterial function and encouraging MVPA alongside strategies to enhance healthy muscle development, especially in females, while recognizing possible remodeling effects on carotid structure with higher MVPA.

Higher sedentary time was paradoxically associated with lower cIMT and not with cfPWV; middle tertile ST related to higher carotid elasticity. Higher LPA and the highest LPA tertile were associated with lower cfPWV, and moderate LPA tertile with higher carotid elasticity. Higher MVPA and meeting ≥60 min/day MVPA were associated with higher carotid elasticity and paradoxically with higher cIMT, with lean mass as the strongest mediator of the ST–cIMT and MVPA–cIMT/elasticity relationships; insulin resistance partially mediated MVPA–elasticity. Total fat mass, trunk fat mass, and BP did not significantly mediate movement behavior–vascular associations. Preventing subclinical arteriosclerosis and atherosclerosis in mid-to-late adolescence may focus on improving LPA and MVPA and increasing muscle mass, particularly in females. Future research should further elucidate mechanistic pathways and long-term implications of MVPA-related vascular remodeling.

Potential misclassification between ST and LPA due to 60-second accelerometer epochs; absence of movement behavior measures at age 17 (when vascular outcomes were assessed), though time interval was adjusted; limited reliability from representing approximately 30 hours of wear over 7 days despite objective assessment; observational design limits causal inference; potential low power in hypertensive subgroup analyses; predominantly Caucasian sample limits generalizability; possible under/overestimation of movement behaviors inherent to accelerometry processing choices.