Associations of Vigorous-Intensity Physical Activity with Biomarkers in Youth

Abstract

Introduction: Physical activity (PA) conveys known cardiometabolic benefits to youth, but the contribution of vigorous-intensity PA (VPA) to these benefits is unknown. Therefore, we sought to determine (a) the associations between VPA and cardiometabolic biomarkers independent of moderate-intensity PA (MPA) and time sedentary and (b) the accelerometer cut point that best represents the threshold for health-promoting VPA in youth.

Methods: Data from the International Children's Accelerometry Database (ICAD) were analyzed in 2015. The relationship between cardiometabolic biomarkers and four categories of VPA estimated via three sets of cut points were examined using isotemporal substitution quantile regression modeling at the 10th, 25th, 50th, 75th, and 90th percentile of the distribution of each biomarker, separately. Age, sex, accelerometer wear time, sedentary time, and MPA were controlled for while allowing substitution for light-intensity PA. Data from 11,588 youth (4-18 yr) from 11 ICAD studies (collected 1998-2009) were analyzed.

Results: Only 32 of 360 significant associations were observed. Significant, negative relationships were observed for VPA with waist circumference and insulin. Replacing light-intensity PA with VPA (corresponding to at the 25th to 90th percentiles of VPA) was associated with 0.67 (-1.33 to -0.01; P = 0.048) to 7.30 cm (-11.01 to -3.58; P < 0.001) lower waist circumference using Evenson and ICAD cut points (i.e., higher counts per minute). VPA levels were associated with 12.60 (-21.28 to -3.92; P = 0.004) to 27.03 pmol·L (-45.03 to -9.03; P = 0.003) lower insulin levels at the 75th to 90th percentiles using Evenson and ICAD cut points when substituted for light PA.

Conclusions: Substituting light PA with VPA was inversely associated with waist circumference and insulin. However, VPA was inconsistently related to the remaining biomarkers after controlling for time sedentary and MPA.

Figure

Combination of quantile regression model coefficients and standard errors for each risk factor across the 11 studies for each of the three sets of accelerometer cutpoints.

Figure

Combination of quantile regression model coefficients and standard errors for each risk factor across the 11 studies for each of the three sets of accelerometer cutpoints.

Figure

Combination of quantile regression model coefficients and standard errors for each risk factor across the 11 studies for each of the three sets of accelerometer cutpoints.

Figure

Combination of quantile regression model coefficients and standard errors for each risk factor across the 11 studies for each of the three sets of accelerometer cutpoints.