Physical activity (PA) has many health benefits and lack of PA is associated with higher risk of a range of chronic non-communicable diseases. Globally, more than 7% of all-cause mortality could be attributable to physical inactivity, (British Journal of Sports Medicine 2022) with the proportions varying for different diseases. Based chiefly on data from Western studies, the WHO 2020 Guidelines suggest that to maintain good health adults should engage in at least 150–300 min per week of moderate physical activity (or 75–150 min per week of vigorous PA) (British Journal of Sports Medicine 2020). A decade of research in CKB has provided new evidence about the levels, patterns and health effects of physical activity on different disease outcome in Chinese adults.
In CKB the mean total physical activity was much higher than that seen in Western populations and it was mainly from occupation (62%) and household chores (26%), but little from active-recreation (4%).
Overall, women in CKB had lower total physical activity but much higher household activity than men. Older participants had a lower level of occupational activity but a higher level of household and active-recreational activity, particularly after retirement. (BMC Public Health 2014). Evidence from the CKB also found that physical activity was about one-third lower among professionals than among factory workers, with intermediate levels for other occupational categories. Unsurprisingly, greater physical activity was associated with lower body mass index, smaller waist circumference, and lower percentage points of body fat. (Am J Clin Nutr 2013).
Previous studies have reported different health effects for leisure time physical activity (LTPA) and occupational physical activity, with potential health benefits for the former but not the latter for cardiovascular disease (CVD), a phenomenon known as the PA paradox. Detailed analyses of CKB data found that total physical activity was inversely associated with the risk of major CVD, with no evidence of any threshold throughout the range of total physical activity studied. It was estimated that approximately 1 hour of brisk walking per day or the equivalent, was associated with a 6% lower overall CVD risk, with similar or somewhat stronger associations with particular types of CVD such as ischemic stroke, intracerebral haemorrhage, and CVD death. The strength of the associations was similar and independent of each other for occupational and non-occupational physical activity. These comprehensive analyses in CKB provided no supportive evidence for the PA paradox and demonstrated the health benefits of both occupational and non-occupational physical activity for major CVD (JAMA Cardiol. 2017).
Domain-specific PA is generally under researched. Overall, 47.2% of CKB participants reported non-active commuting (e.g. using public transport or driving), 13.4% reported working at or near home, 20.1% reported walking, and 19.4% reported cycling to work. We found that walking and cycling were associated with a lower risk of ischaemic heart disease than non-active commuting. Cycling was associated with a lower risk of ischemic stroke and the associations of commuting mode with major CVD were consistent among men and women and across different levels of other physical activity (JAHA 2019).
Unlike most previous studies, CKB has also assessed the associations of physical activity and sedentary behaviour with disease risks independently and combined. In CKB higher levels of physical activity were associated with lower risks of diabetes, whereas the converse was true for higher levels of sedentary leisure-time (SLT). Overall, highest versus lowest PA group was associated with 16% lower risk, while highest versus lowest SLT group was associated with 20% higher risk. These associations of PA and SLT with risks of diabetes were independent of each other, but were more modest after additional adjustment for body mass index. While PA and SLT may be important risk factors for diabetes, these associations appear to be chiefly mediated through adiposity (BMJ Open Diabetes Research and Care 2019).
Hepatobiliary disease, which affects the liver, gallbladder and bile ducts, accounted for 2.4 million deaths globally in 2017, with about 50% of deaths from liver cancer and 15% from cirrhosis in China. We found that high total physical activity, particularly occupational physical activity, was inversely associated with risk of major hepatobiliary cancers and diseases, including non-alcoholic fatty liver disease (NAFLD), cirrhosis and certain types of cancer (British Journal of Sports Medicine 2021). Overall, when compared with the lowest PA group, the highest group had approximately 25% lower risk of hepatobiliary diseases, with similar associations for NAFLD, cirrhosis and viral hepatitis.
Regular physical activity is associated with improvements in mental health, quality of life and well-being. Overall, 12.9% of male and 17.8% of female participants in CKB reported having insomnia symptoms. After adjustment for potential confounders, a moderate to high level of overall activity was associated with between 6% and 16% lower risks of difficulties in falling asleep or maintaining sleep and daytime dysfunction in both sexes. For each domain of physical activity, similar associations were identified for occupational, household and leisure-time activities in women but not men. Commuting-related activity, however, was associated with higher risks of difficulties in initiating or maintaining sleep and any insomnia symptoms in both sexes (Journal of Sleep Research 2017).
We used NMR-based metabolomic biomarkers data in CKB to explore the mechanisms underlying the relationship between physical activity and sedentary leisure time with occlusive CVD. We found that PA and SLT were associated with more than 100 metabolic markers, with patterns of associations generally mirroring each other. Physical activity was inversely associated with blood levels of lipids amino acids, inflammatory markers, glucose, and fatty acids. In general, associations of PA and SLT with specific metabolic markers were directionally consistent with the associations of these metabolic markers with occlusive CVD risk. Overall, these metabolic markers potentially explained approximately 70% of the inverse associations of PA and about 50% of the positive associations of SLT with occlusive CVD (CIRCGEN 2019).
Impact of research
Research based on CKB data has provided important local evidence on the health benefits of increasing overall physical activity levels. Several key CKB findings were incorporated into policy guidelines for prevention of chronic diseases in China (Chinese Journal of Public Health 2022). In response to the rising burden of chronic diseases, the Chinese government has developed a Healthy China 2030 Plan, with promotion of physical activity listed as one of the key actionable areas (J Sport Health Sci. 2017).
Self-reports of physical activity and sedentary behaviour are prone to recall bias and measurement error. Although more expensive and time-consuming, accelerometers are reliable and valid objective instruments for measuring PA. Our future work will examine the wrist-worn accelerometer data recently collected in the third resurvey of 25,000 CKB participants. These detailed measurements will enable more sophisticated and reliable analyses to identify the shape and strength of associations with disease risks. It will also enable, in conjunction with new biomarker data, discovery of likely mechanisms of action related to physical activity.