To investigate the effect of body weight, waist circumference and their changes on all-cause and cardiovascular mortality.

A nationwide population-based cohort study

627 community-dwelling older adults.

Participants were interviewed for demographic and anthropometric data collected. Blood were drawn for testing biochemistry data. Central obesity was defined as waist circumference is greater than 80 cm for women and 90 cm for men. Obesity, overweight, normal and underweight were defined as BMI ≥27 kg/m2, ≥24 kg/m2,18.5-24 kg/m2 and <18.5 kg/m2. Cox proportion hazard model was used to explore the impact of body weight and its change on mortality.

The distribution of weight changes and mortality was right skewed, but U-shape of waist change for all-cause mortality was observed. Compared to normal BMI at baseline, the association between underweight (HR: 1.7, 95% CI: 0.7-4.0), overweight (HR:0.7, 95% CI:0.4-1.2) and obesity (HR:1.3,95% CI:0.8-2.3) showed insignificantly associated with all-cause mortality. The HR of those weight loss >5% (HR: 1.7, 95% CI: 1.1-2.8) and waist decrease >5% (HR: 1.7, 95% CI: 1.0-2.8) were higher than those of stable weight/waist +/- 5% over a 6-year period. Compared to those stable weight/waist, the mortality risk was similar in those of weight gain or waist increase (HR 0.7,95%CI: 0.4-1.5 and HR:0.9, 95%CI:0.4-1.6).

Weight loss and waist decrease were significantly associated with long-term mortality risk, a life-course approach for body weight management is needed to pursuit the most optimal health benefits for the middle-aged and older adults.