Long-term time in target range for body mass index (BMI) and diabetes incidence: insights from CHARLS

Nutrition & Diabetes ARTICLE www.nature.com/nutd OPEN Long-term time in target range for body mass index (BMI) and diabetes incidence: insights from CHARLS ✉ Chao Chu1 , Yan Zhang2, Fuxue Deng2 and Lisha Zhang 2✉ 1234567890();,: © The Author(s) 2025 BACKGROUND: Body mass index (BMI) is strongly associated with the development of type 2 diabetes. However, the association between long-term time in target range (TTR) for BMI and the incidence of new-onset diabetes remains unclear. METHODS AND RESULTS: This study utilized a non-diabetic population aged 45 years or older from the China Health and Retirement Longitudinal Study (CHARLS). BMI-TTR was assessed in Waves 1, 2, and 3 over a 5-year period, with the target range defined as 18.5 kg/m² ≤ BMI < 23 kg/m². New-onset diabetes in Waves 2, 3, and 4 over a 6-year follow-up served as the study endpoint. After applying exclusion criteria, 6662 participants (3143 men and 3519 women; mean age 58.93 ± 8.85 years) were enrolled. Participants were categorized into four groups (TTR1–TTR4) based on the number of times BMI was within the target range (0–3 times). The risk of new-onset diabetes decreased progressively with increasing BMI-TTR during follow-up. Compared with the TTR1 group, participants in the TTR4 group exhibited a significantly lower risk of diabetes (adjusted HR: 0.577, 95% CI: 0.463–0.720, P < 0.001), even after adjusting for baseline BMI (adjusted HR: 0.685, 95% CI: 0.537–0.872, P = 0.002). This effect was even more pronounced in female subgroup and in individuals aged under 60 years of age. CONCLUSION: In adults aged 45 years or older, regardless of baseline BMI, maintaining BMI within the target range over time was associated with a reduced risk of new-onset diabetes, particularly among women and individuals under 60 years of age. These findings highlight the importance of long-term weight management in diabetes prevention. Nutrition and Diabetes (2025)15:46 ; https://doi.org/10.1038/s41387-025-00404-8 BACKGROUND Type 2 diabetes mellitus (T2DM) is a major public health concern and a leading cause of mortality and disability worldwide. Studies have shown that body mass index (BMI) is a commonly used measure of obesity and that obese individuals (as measured by BMI) have a significantly higher risk of T2DM [1, 2]. This association is particularly pronounced among the elderly, especially those over 50 years of age [3]. Several studies have indicated that a normal BMI, not too low nor too high, is significantly associated with a reduced risk of T2 DM, cardiovascular disease events, and death [4–6]. Time in target range (TTR) metrics, such as those for blood glucose and blood pressure, have been employed to predict complications and clinical outcomes in diabetes and hypertension [7–12]. Elevated BMI has also been independently associated with an increased incidence of cardiovascular disease incidence and mortality [13, 14]. However, the association between BMI-TTR and the risk of new-onset diabetes remains underexplored. In this study, we utilized the China Health and Retirement Longitudinal Study (CHARLS), which initially recruited a middleaged and older cohort aged 45 years and older and has followed them over time, to investigate the association between BMI-TTR and risk of the long-term risk of diabetes in a middle-aged and elderly population. PARTICIPANTS AND METHODS Study design and participants CHARLS is a longitudinal cohort study that collects high-quality microdata on households and individuals aged 45 years and older in China, and the study is ongoing. For a detailed description of CHARLS, please refer to the relevant literature and its official website [15, 16]. Participants included in this analysis had BMI measurements from CHARLS Wave 1 (2011), Wave 2 (2013), and Wave 3 (2015). Individuals under 45 years of age, those with diabetes mellitus (DM) at baseline, and those with a fasting blood glucose ≥7 mmol/L or glycated hemoglobin ≥6.5% at baseline were excluded. Ultimately, a total of 6662 participants were included in the BMI-TTR analysis (Flow Diagram was shown in Fig. 1). The study was approved by the Ethical Review Committee of Peking University. The studies were in accordance with the STROBE Statement, and informed consent was obtained from each participant. Assessment of BMI and definition of BMI-TTR BMI was calculated as weight (kg) divided by height squared (m²). The target range for BMI was defined as 18.5 kg/m² ≤ BMI < 23 kg/m², based on World Health Organization (WHO) recommendations for Asian populations. Participants were stratified into four groups (TTR1–TTR4) based on the number of times their BMI fell within the target range (0–3 times). The TTR1–TTR4 categories are mutually exclusive. The TTR1 group consists of participants whose BMI was outside the normal range in all three Waves. The TTR2 group refers to participants whose BMI fell within the normal range in any one of the three Waves. The TTR3 group refers to participants whose BMI fell within the normal range in any two of the three Waves. The 1 Department of Cardiovascular Medicine, the First Affiliated Hospital, Xi’an Jiaotong University, Xi’an, Shaanxi, PR China. 2Department of Cardiovascular Medicine, the Second Affiliated Hospital, Xi’an Jiaotong University, Xi’an, Shaanxi, PR China. ✉email: ; Received: 7 May 2025 Revised: 6 October 2025 Accepted: 16 October 2025 C. Chu et al. 2 Fig. 1 Flow diagram. Flow diagram showing the selection of the study population. TTR4 group includes participants whose BMI remained within the normal range across all three Waves. Incident diabetes New-onset diabetes, defined as self-reported physician-diagnosed diabetes, was assessed in Waves 2 (2013), 3 (2015), and 4 (2018) over a 6-year period. Statistical analysis Statistical analyses were performed using SPSS 26.0 (SPSS Inc. Chicago, IL) and R software (version 4.3.3, R Foundation for Statistical Computing, Austria). Categorical variables were expressed as percentages, and continuous variables as means ± SD. Group comparisons were performed using Student t-tests, ANOVA, nonparametric tests, Chi-square tests, or Fisher’s exact tests, as appropriate. Cox regression models and binary logistic regression models were employed to calculate hazard ratios (HR) and odds ratios (OR) with 95% confidence intervals (CI). Kaplan–Meier curves illustrated the incidence of outcomes. Multivariate Cox models and binary logistic regression models adjusted for covariates including age, gender, hypertension, smoking, and other relevant factors. The COX regression models included the following: Model 1: Age, gender, hypertension, smoking, systolic blood pressure (SBP), heart rate (HR), hyperlipidemia, physical activity, marital status, education, high-density lipoprotein cholesterol (HDL-C), glycated hemoglobin A1c (HbA1c), high sensitivity C-reactive protein (hsCRP), estimated glomerular filtration rate (eGFR), and BMI-TTR. Model 2: Age, gender, hypertension, smo (...truncated)