Association of Leisure-Time Physical Activity With Mortality From Diabetes Among Older Chinese Adults: A Multicity Prospective Cohort Analysis

Background Physical activity is associated with lower mortality, but whether the frequency of leisure-time physical activity (LTPA) is associated with mortality from diabetes among older adults has not been well characterized. Clarifying this association may inform the interpretation of routinely collected exercise information in community health management in China. Methods We conducted a prospective cohort analysis of 1,565,890 Chinese adults aged 65 years or older enrolled between 2017 and 2019 at three study sites (Luzhou, Shenzhen, and Zaozhuang) of the Multi-city Elderly Health Examination Cohort Study (MEHECS). LTPA frequency was categorized as no LTPA, occasional LTPA, LTPA at least once weekly but not daily, and daily LTPA. Participants were followed through December 31, 2024. Cause-specific Cox models estimated hazard ratios (HRs) and 95% confidence intervals (CIs) for mortality from diabetes, with deaths from other causes censored; all-cause mortality was secondary. Models adjusted for demographic, socioeconomic, lifestyle, anthropometric, chronic-disease, and city variables. Fine–Gray regression and restriction to participants with more than 2 years of follow-up were performed as sensitivity analyses; models for mortality from diabetes were additionally repeated after excluding participants with baseline diabetes. Results During a median follow-up of 5.06 years (interquartile range, 3.30–6.73 years), 5,584 deaths from diabetes and 185,815 all-cause deaths occurred. Compared with no LTPA, the fully adjusted HRs for mortality from diabetes were 0.86 (95% CI, 0.74–1.01) for occasional LTPA, 0.83 (0.73–0.94) for LTPA at least once weekly but not daily, and 0.67 (0.62–0.72) for daily LTPA. Corresponding HRs for all-cause mortality were 0.85 (0.82–0.87), 0.81 (0.79–0.82), and 0.67 (0.66–0.68). Among participants reporting LTPA, longer session duration and more years of regular LTPA were associated with lower risks of mortality from diabetes and all-cause mortality after adjustment for LTPA frequency. The overall pattern was broadly similar in competing-risk and early-follow-up analyses; daily LTPA remained inversely associated with mortality from diabetes in all sensitivity analyses. Conclusions LTPA frequency derived from routinely collected community health-examination data was prospectively associated with lower mortality from diabetes among older Chinese adults, with the most consistent association observed for daily LTPA. More detailed and repeated activity assessment is needed to clarify the independent contributions of frequency, duration, intensity, and total activity volume.