Objective To examine the association between physical activity and risk of all-cause mortality and related mediators in middle-aged and elderly people in China to provide evidence for reducing mortality in the population.

Methods Data were collected from 5 727 individuals aged ≥ 45 years who completed baseline and 3 follow-up surveys in the China Health and Retirement Longitudinal Study (CHARLS) from 2011 to 2018. We used multivariable Cox proportional hazards regression models and restricted cubic spline models to examine the association between physical activity levels and risk of all-cause mortality. In addition, we conducted a mediation analysis based on body mass index (BMI), blood pressure, and blood biochemical indicators collected during follow-up.

Results Of all the individuals analyzed, 1 813 (31.66%) were classified as having physical activity (measured in metabolic equivalents METs-min/week) in the lowest quartile (<1 720.95), and the number (proportion) of the individuals with physical activity in the second quartile (1 720.95 – 5 543.99), third quartile (5 544.00 – 12 263.99), and highest quartile (≥ 12 264.00) were 1 082 (18.89%), 1374 (23.99%), and 1458 (25.46%), respectively. During the follow-up, a total of 509 deaths were observed and the all-cause mortality rate was 8.92%, with the number of all-cause deaths (mortality rate) being 271 (14.99%), 85 (7.88%), 80 (5.89%), and 73 (5.01%) among those with physical activity in the lowest, second, third, and the highest quartiles of METs-min/week, respectively. All-cause mortality showed a significant downward trend with increasing levels of physical activity among the individuals (χ²_trend = 653.81, P _trend < 0.001). After adjustment for sex, age, education level, marital status, place of residence, smoking, alcohol consumption, BMI, systolic blood pressure (SBP), diastolic blood pressure (DBP), white blood cells (WBC), mean corpuscular volume (MCV), platelets (PLT), hemoglobin (Hb), hematocrit (HCT), C-reactive protein (CRP), glucose (GLU), hemoglobin A1c (HbA1c), total cholesterol (TC), triglycerides (TG), low-density lipoprotein (LDL), high-density lipoprotein (HDL), blood urea nitrogen (BUN), creatinine, uric acid (UA), and cystatin C (CysC), multivariate Cox proportional hazards regression analysis showed that the hazard ratio for all-cause mortality for individuals in the second, third, and highest quartiles of METs-min/week was 0.62 (95% confidence interval 95%CI: 0.49 – 0.80), 0.52 (95%CI: 0.40 – 0.67), and 0.46 (95%CI: 0.35 – 0.61) times, respectively, compared with those in the lowest quartile. Furthermore, the results of the restricted cubic spline model showed that there was a non-linear dose-response relationship between physical activity level and risk of all-cause mortality among individuals. This relationship followed an L-shaped curve, indicating that the risk of all-cause mortality gradually decreased with increasing levels of physical activity. However, after a certain threshold of physical activity (16 000 METs-min/week), the risk of all-cause mortality no longer decreased significantly with increasing physical activity. In addition, the mediation analysis suggested that BMI and CRP might play a mediating role in the association between physical activity and risk of all-cause mortality, with the percentage of mediation effects among individuals in the highest quartile of physical activity being 37.22% and 39.60%, respectively.

Conclusion Increasing physical activity levels may reduce the risk of all-cause mortality among middle-aged and older adults in China, and appropriate physical activity may provide greater health benefits for this population. Reduced BMI and inflammatory responses in the body may be potential mechanisms for the association between physical activity levels and risk of all-cause mortality.