Résumé
Objective: To explore the effect of BMI and waist circumference on diabetes of adults. Methods: After excluding participants with heart disease, stroke, cancer and diabetes at baseline study, 53 916 people aged 30-79 in the China Kadoorie Biobank (CKB) study from Tongxiang city of Zhejiang province were recruited. Cox regression model was used to estimate the hazards ratios (HR) for the associations of baseline BMI and waist circumference with incident diabetes. Results: Among 391 512 person-years of the follow-up program between 2004 and 2013 (median 7.26 years), a total of 944 men and 1 643 women were diagnosed as having diabetes. Compared to those with normal weight, after adjusting for known or potential factors, HR of both overweight and obesity in men for incident diabetes appeared as 2.72 (95%CI: 2.47-2.99) and 6.27 (95%CI: 5.33-7.36), respectively. The corresponding figures in women were 2.19 (95%CI: 2.04-2.36) and 3.78 (95%CI: 3.36-4.26). Compared to those with normal waist circumference, after adjusting for known or potential factors, HR of Ⅰgrade andⅡgrade in men for diabetes were 2.56 (95%CI: 2.22-2.95) and 4.66 (95%CI: 4.14-5.24), respectively. The corresponding figures in women were 1.99 (95%CI: 1.80-2.21) and 3.16 (95%CI: 2.90-3.44), respectively. Conclusions: Overweight, obesity and central obesity were all associated with the increased incident of diabetes. Strategies on diabetes prevention should include not only losing weight, but reducing waist circumference as well.