Association of change in cognitive function from early adulthood to middle-age with risk of cause-specific mortality

Background: Studies with single baseline measurements of cognitive function consistently reveal inverse relationships with mortality risk. The impact of change in functioning, particularly from early in the life course, which may offer additional insights into causality, has not, to the best of our knowledge, been tested.  
Aims: To examine the association of change in cognition between late adolescence and middle-age with cause-specific mortality using data from a prospective cohort study. 
Methods: The analytic sample consisted of 4289 US male former military personnel who were administered the Army General Technical Test in early adulthood (mean age 20.4 yr.) and again in middle-age (mean age 38.3 yr.).
Results: A 15 year period of mortality surveillance subsequent to the second phase of cognitive testing gave rise to 237 deaths. Following adjustment for age, a ten unit increase in cognitive function was related to a reduced risk of death from all-causes (hazard ratio; 95% confidence interval: 0.84; 0.75, 0.93) and cardiovascular disease (0.78; 0.64, 0.95) but not from all cancers (1.14; 0.88, 1.47) nor injury (1.02; 0.81, 1.29). Adjustment for markers of socioeconomic status in middle-age resulted in marked attenuation in the magnitude of these associations and statistical significance at conventional levels was lost in all analyses. 
Conclusions: Increases in cognitive function earlier in the life course were associated with lower mortality risk, and these effects were mediated by socioeconomic status in the present study.