Association of physical activity with insulin resistance: A cross-sectional mediation analysis of the systemic immune-inflammation index in NHANES.

Abstract

Higher levels of physical activity (PA) are associated with lower insulin resistance (IR), but the underlying mechanisms remain incompletely understood. Systemic inflammation may represent one pathway linking PA to IR. This cross-sectional study included 9088 adults aged ≥ 18 years from the 2005 to 2016 National Health and Nutrition Examination Survey. IR was assessed using the triglyceride-glucose index, total PA was calculated as metabolic equivalent minutes per week (MET-min/week), and the systemic immune-inflammation index (SII) was used as an inflammatory marker. Survey-weighted multivariable linear regression models were used to examine the association between PA and IR. Restricted cubic spline analyses were used to assess potential nonlinearity, and mediation analysis was performed to evaluate the mediating role of SII. Total PA was inversely associated with IR after multivariable adjustment (β [95% CI] = -0.089 [-0.127 to -0.052], P < .001). Restricted cubic spline analysis indicated a nonlinear association, with a clear inverse relationship between PA and IR at low-to-moderate PA levels and greater uncertainty beyond approximately 1922 metabolic equivalents of task-min/week. Mediation analysis showed that SII partially mediated the association between PA and IR (β [95% CI] = -0.005 [-0.009--0.003], P < .01). Sensitivity analysis using multiple imputation for missing covariates showed a consistent inverse association between PA and IR after full adjustment. In a nationally representative US sample, higher PA levels were associated with lower IR, and systemic inflammation accounted for a small proportion of this relationship. The PA range associated with lower IR may help inform exercise-based strategies for the prevention and management of metabolic disorders.