Latent profile analysis of physical activity and sedentary behavior with mortality risk: A 15-year follow-up

By exploring multiple characteristics of physical activity and sedentary behavior (SB), different physical activity profiles could be obtained, which may be beneficial for health and targeted physical activity interventions. The aim of this study was to identify distinct physical activity profiles based on accelerometer-derived activity characteristics and to determine whether these profiles are associated with all-cause mortality. Eight hundred fifty-one participants (56% women, mean age: 53 years) provided objectively assessed physical activity data using an ActiGraph accelerometer and were followed for 15 years. Physical activity profiles were determined using latent profile analyses of 14 derived activity variables, resulting in that three profiles were identified: "Low Active" (n = 147), "Average Active" (n = 397), and "High Active" (n = 307). "Low Active" was characterized by participants with low absolute, relative, and limited variation of time spent in physical activity, and high time spent in SB. "Average Active" had the most balanced movement behavior with values close to the mean for all activity variables. "High Active" was characterized by participants with high absolute, relative, and great variation of time spent in physical activity. Overall, a potentially non-linear pattern between multiple activity variables and all-cause mortality was found as "Low Active" was significantly (P < .05) positively associated with all-cause mortality, and no difference in mortality risk was found between "High Active" and "Average Active." Our data suggest that day-to-day variation in SB is not associated with all-cause mortality. The important message is to keep the overall time spent in SB low and replace this behavior with physical activity.