2107-LB: The Impact of Obesity Phenotypes and Short-Term Dynamic Transitions on Diabetes Risk in Chinese Middle-Aged and Elderly Populations—A Longitudinal Study Based on the CHARLS Cohort

Introduction and Objective: To examine the longitudinal effects of baseline obesity phenotypes and their transitions on diabetes risk in Chinese adults aged ≥45 years, with a particular focus on the role of abdominal obesity in normal-weight individuals. Methods: Based on data from the CHARLS cohort, this study included 4,543 non-underweight participants free of diabetes at both baseline and the 3-year follow-up. According to Chinese BMI and abdominal obesity criteria, baseline phenotypes were categorized into: NW-NAO (normal weight/non-AO), HB-NAO (high BMI/non-AO), NW-AO (normal weight/AO), and HB-AO (high BMI/AO). Multivariable Cox proportional hazards models were employed to assess the associations of baseline phenotypes and their trajectories from baseline to the second year with incident diabetes risk after the third year, adjusting for confounders including age, sex, smoking, alcohol consumption, and baseline glycated hemoglobin (HbA1c). Results: During a median follow-up of 8.42 years, 737 incident diabetes cases were recorded. Compared with the NW-NAO group, the baseline HB-NAO, NW-AO, and HB-AO groups exhibited higher risks with HR 1.44(1.08,1.91), HR 1.65(1.29,2.12), and HR 1.73(1.43,2.09), respectively. Longitudinal analysis demonstrated that the HB-NAO group maintaining status, transitioning to NW-AO through waist circumference gain, or progressing to HB-AO had 58%, 3.3-fold, and 68% elevated risks, respectively. The NW-AO group persisting with abdominal obesity or developing HB-AO via BMI increase showed 106% and 61% increased risks, respectively. The HB-AO group maintaining the phenotype had a 91% higher risk. No protective effects were observed with 2-year phenotype improvements. Conclusion: Abdominal obesity independently predicts diabetes risk, even in normal-weight individuals. Sustained or aggravated obesity phenotypes amplify diabetes susceptibility, whereas short-term improvements fail to mitigate risks. Disclosure Y. Liu: None. J. Zhang: None. S. Gao: None. M. Han: None. L. Zhang: None. X. Jia: None. M. Xie: None. T. Wang: None. H. Zhu: None. M. Li: None. Z. Song: None.