Physical activity and sedentary behavior

Abstract

The physical activity (PA) and sedentary behavior (SED) levels of college students may be related to their cardiometabolic health. This study aimed to assess the relationships between measures of PA and SED with each other and a criterion measure and their respective associations with cardiometabolic risk indicators among college freshmen. Students (n = 402; age = 18.3 ± 0.5 years; 68.7% female) completed the Godin Leisure Time Questionnaire, the IPAQ, the GPAQ, and a television viewing question, and wore a New Lifestyles 1000 (NL) accelerometer. A subgroup (n = 54) also wore an activPAL. Cardiometabolic risk was indicated by HOMA, C-reactive protein (CRP), android body fat (BF%), and the Pathobiological Determinants of Atherosclerosis in Youth (PDAY) risk score. ActivPAL steps were associated with Godin vigorous activity, IPAQ travel activity, and NL steps, and aerobic steps were associated with Godin total vigorous activity and NL steps and activity minutes [model R2 = 0.76 and 0.68, respectively, p < 0.001]. ActivPAL sitting time was independently associated with television time and NL steps [model R2 = 0.14, p = 0.02]. NL steps and Godin total activity independently predicted BF% for PA measures [model R2 = 0.35 and 0.34, respectively, p < 0.001]. Adjusted for gender, significant associations between HOMA and the PA measures of NL steps and Godin total activity were seen (highest quartile of PA had 50% reduced prevalence compared to lowest group, p < 0.10). Television was associated with CRP (91% increased prevalence in highest third compared to lowest, p = 0.05). Regression analyses did not retain any multiple measure models across outcomes. Latent variables did not have a greater predictive ability of activPAL measures or cardiometabolic outcomes than field measures. No interaction effect was present between levels of PA and SED or with BF%. The findings suggest that NL steps and Godin measures were mostly strongly associated with activPAL criterion estimates and with cardiometabolic risk in college students. Latent variables were not superior to individual measures. Future work should seek methods of combining complementary measures to improve predictive ability in college students.