Background and objective: Physical activity (PA) confers protection to individuals from the risk of death. However, in the very old, the dose-response relationship between PA and all-cause mortality and the possible biological mediators of this association are less known. We investigated whether PA predicts 6-year all-cause mortality and what biomarkers mediate the association. Design: Prospective cohort data from the Tokyo Oldest Old Survey on Total Health study. Setting: Community-dwelling population. Participants: A total of 441 women and men aged over 85 years. Measurements: Questionnaire-based PA was assessed at baseline and 3-year and 6-year follow-up visits. Survival status was confirmed up to the 6-year follow-up visit (153 deaths, 34.7%). Data of plasma albumin, cholinesterase, NT-proBNP, interleukin-6, cystatin C, and HbA1c levels were collected. For mediation analysis for survival analysis, we used the baseline PA and biomarkers with Weibull distribution accelerated failure time model and linear regression model adjusted for age, sex, body mass index, smoking, education level, and Mini-Mental State Examination. Results: A curvilinear relationship was observed in the association between baseline PA and all-cause mortality. Compared to the inactive (0 METs*h/week), light amount of PA was associated with a lower risk of mortality. Compared to the highest tertile of PA (11.2 METs*h/week), higher PA did not reduce the risk of death. Circulation levels of albumin and cholinesterase mediated the association between baseline PA and all-cause mortality (proportion mediated, 54%, both; p < 0.05). Conclusions: Compared to completely inactive, light PA reduces the risk of all-cause mortality in the very old population. Mediation analysis suggests that protein synthesis in the liver may mediate the association between PA and all-cause mortality. Further studies are needed to understand the underlying association between PA, nutrition, and death.
ASJC Scopus subject areas