Longitudinal Assessment of Hip Fractures-Diabetes Causality in Aging Populations: Evidence from the China Health and Retirement Longitudinal Study Cohort

Objective: To examine the longitudinal impact of hip fractures on the risk of diabetes in middle-aged and older individuals, and investigate the mediating effects of depression, sleep duration, and self-medication with traditional herbal medicines and tonic health supplements. Methods: Data from the China Health and Retirement Longitudinal Study (10,280 participants, aged 45+) were analyzed over a 7-year follow-up. Subgroup analyses were conducted by gender, age, and body mass index. After adjusting for relevant confounders, the Cox proportional hazards model was used to assessed the association between hip fractures and diabetes risk. Linear regression models were employed to conduct mediation analyses. Structural equation model was applied for sensitivity analysis. Results: At the 2011 baseline, 152 participants (1.48%) had a history of hip fractures. By 2018, 559 participants (5.44%) developed diabetes, including 10 (1.79%) individuals with prior hip fractures. Hip fractures were significantly associated with increased diabetes risk in males (HR = 2.43; 95% CI: 1.19, 4.96; P = 0.01) and obese participants (HR = 3.77; 95% CI: 1.42, 10.02; P < 0.01). No significant causal association was observed between hip fractures and diabetes in the overall population or in subgroups defined by female, age, or BMI < 28. Self-medication with herbal medicines and tonic supplements may partially explain the increased diabetes risk among obese individuals with hip fractures, while depression and sleep duration showed no significant mediating effects. Conclusion: Hip fractures were significantly associated with an increased risk of diabetes among middle-aged and older males, as well as individuals with obesity. To mitigate this risk, it was recommended to standardize the use of traditional herbal medicines and tonic health supplements under the guidance of healthcare professionals.