Dose-response relationship between physical activity and visceral fat mass: a cross-sectional study based on NHANES 2011–2018

Excess visceral fat mass (VFM) is a key risk factor for metabolic diseases, and physical activity (PA) is a modifiable behavior associated with lower VFM. This cross-sectional study investigates the dose-response relationship between PA and VFM, with a focus on differences by sex and race/ethnicity, using data from the National Health and Nutrition Examination Survey (NHANES) 2011-2018. Among 9,111 adults aged 20-59 years, PA was measured as metabolic equivalent of task (MET) minutes per week via the Global Physical Activity Questionnaire (GPAQ) and categorized into quintiles (Q1: 0-280, Q2: 281-1,680, Q3: 1,681-5,400, Q4: 5,401 - 11,020, Q5: >11,020 MET-min/week). VFM was assessed using dual-energy X-ray absorptiometry (DXA). Weighted linear regression and restricted cubic spline (RCS) models were employed to examine associations, adjusting for demographic (age, sex, race/ethnicity), lifestyle (smoking, alcohol, caloric intake), and health-related (BMI, diabetes, hypertension) covariates. Higher PA levels were associated with lower VFM ratios, with the highest quintile (Q5) having a 12% lower ratio compared to the lowest (Q1) (Ratio = 0.88, 95% CI: 0.85-0.91, P < 0.001) after full adjustment.RCS analyses revealed sex-specific patterns: males exhibited a monotonic decrease in VFM up to 50,000 MET-min/week, while females showed a U-shaped association, with the lowest VFM at ~ 6,000 MET-min/week, suggesting moderate PA (~ 5,000-7,000 MET-min/week) as optimal. Significant interactions by sex (P_interaction = 0.008) and race/ethnicity (P_interaction < 0.001) highlighted heterogeneity: Non-Hispanic Whites showed a consistent VFM decline, while Mexican Americans exhibited a weaker association with An inverted U-shape, peaking at 1,000-3,000 MET-min/week. These findings underscore the need for tailored PA recommendations by sex and race/ethnicity. The cross-sectional design precludes causal inference.