Abstract Funding Acknowledgements Type of funding sources: None. Background Given the crucial role of obesity as a risk factor for progression of cardiovascular disease (CVD), various guidelines have emphasized the importance of its prevention 1,2. The non-linear relationships of changes in weight, body mass index (BMI), and waist circumference (WC) to all-cause deaths were observed. Evidence on the associations of changes in obesity indices with CVD mortality is scarce. Purposes This study aimed to explore the linear and non-linear relationships of 1- and 3-year changes in obesity indices (i.e., weight, BMI, WC, waist-to-hip ratio (WHR), waist-to-hip ratio (WHRtR)), respectively, to CVD and all-cause mortality. Methods We included adults aged 18 - 65 years from an on-going longitudinal cohort in Taiwan. Multivariable logistics regression models and restricted cubic splines were used to explore the linear and non-linear associations of 1- and 3-year absolute and relative changes in weight, BMI, WC, WHR, and WHtR with CVD and all-cause mortality, respectively. Results A total of 18,024 participants were included in the 1-year change study, and 1,761 and 1,038 cases of all-cause and CVD mortality were reported (follow-up: 6 (range: 2 - 21) years). The 3-year study sample consisted of 12,695 participants, and 1,025 and 931 all-cause and CVD deaths were recorded (follow-up: 6 (range: 2 - 19) years). Results from multivariable logistics regression models showed that the CVD mortality risks were significantly reduced by 1-year absolute and relative increases in weight, BMI, WC, WHR, and WHtR. However, 3-year changes in WC and WHtR did not affect the risk of CVD deaths, while both absolute and relative increases in weight, BMI, and WHR were associated with a lower risk of CVD mortality. For all-cause mortality, any increase in obesity indices was associated with a reduced risk (Table 1). Additionally, restricted cubic splines visually showed U- and L-shaped relationships of changes in obesity indices to CVD and all-cause deaths. Risks of CVD mortality decreased until around 1-year changes of 2.1 cm (2.7%) in WC and 0.01 (2.8%) in WHtR and then increased. Also, below around 3-year change of 2.2 kg (2.5%) in weight, 0.7 kg/m2 (-1.6%) in BMI, 0.03 (3.5%) in WHR, and 0.03 (6.3%) in WHtR, the risks of CVD mortality went down with increases in obesity indices. Above the cut-off points, increased WHR and WHtR were associated with greater risks of CVD mortality. Moreover, an increase was a risk factor for CVD deaths among participants with increased WC of more than approximately 5.5 cm (6.3%) within three years. Similar trends were found among the relationships between changes in obesity indices with all-cause mortality (Figure 1-16). Conclusions We found U- and L-shaped relationships between changes in obesity indices and mortality from CVD and any cause. Risks of CVD and all-cause deaths decreased until a certain value of changes in obesity indices, especially central obesity indices, and then increased.