Physical activity, sedentary leisure-time and risk of incident type 2 diabetes: a prospective study of 512 000 Chinese adults

Objective Aim to examine the independent and joint associations of physical activity (PA) and sedentary leisure-time (SLT) with risk of diabetes and assess the extent to which these associations were mediated by adiposity. Research design and methods The prospective China Kadoorie Biobank recruited similar to 512 000 adults from 10 diverse areas across China. Self-reported PA was estimated based on type, frequency and duration of specific types of PA, covering four domains (occupation, leisure, household and commuting). SLT was defined as hours per day spent watching television, reading or playing card games. Stratified Cox proportional hazards models were used to estimate adjusted HRs (aHRs) for PA and SLT associated with incident diabetes. Analyses were stratified by age-at-risk (5-year intervals), sex and region and adjusted for household income, education, alcohol consumption, smoking, fresh fruit intake, self-reported general health status, family history of diabetes and body mass index (BMI) status. Analyses of total PA, occupational and non-occupational PA and SLT were mutually adjusted for each other, as appropriate. Results After similar to 9 years of follow-up, there were 14 940 incident diabetes cases among 460 736 participants without prior diabetes or cardiovascular diseases at baseline. The mean (SD) age at baseline was 51 (10.6) years, 59% were women and 43% resided in urban areas. Overall, the mean BMI was 23.5 (3.3) kg/m(2), which differed by similar to 0.5 kg/m(2) among individuals in the highest compared with the lowest PA and SLT groups. PA was inversely associated the risk of diabetes 16% (aHR: 0.84, 95% CI 0.81 to 0.88) lower in top than bottom fifth. After further adjustment for BMI this was attenuated to 0.99 (95% CI 0.98 to 1.00). SLT was positively associated with diabetes and each 1 hour per day higher usual level was associated with aHR of 1.13 (95% CI 1.09 to 1.17) for diabetes, attenuated to 1.05 (95% CI 1.01 to 1.09) after further adjustment for BMI. Conclusions Among Chinese adults, higher levels of PA and lower levels of SLT were associated with lower risks of diabetes with no evidence of effect modification by each other. These associations appeared to arise mainly through adiposity.