Commentary on “Association between frailty index and hip fracture among Chinese middle- and old-aged adults in different glycemic status: a prospective study from the China health and retirement longitudinal study (CHARLS)”

Dear Editor, We read with interest the study by Li et al1, exploring the link between the frailty index (FI35) and hip fracture risk under different glycemic conditions using the CHARLS database. The findings highlight a robust linear correlation between frailty and fracture risk, particularly as glycemic control worsens. This study tackled an important public health challenge by bridging frailty, metabolic health, and fracture risk in a representative Chinese population. However, several key factors merit further discussion to contextualize these findings within clinical practice and to guide future research. First, the study omitted the duration of diabetes, a critical confounder in fracture risk assessment. While participants were stratified by baseline glycemic status, the chronicity of diabetes – carrying greater weight in fracture risk assessment than the diagnosis itself – was not adjusted for. Notably, the Manitoba BMD Cohort study established a clear dose-response relationship between diabetes duration and hip fracture risk, with an exceptionally high risk observed in patients with a disease duration exceeding 10 years2. Pathophysiologically, chronic hyperglycemia drives bone strength attrition through oxidative stress, inflammation, and the accumulation of reactive oxygen species and advanced glycation end products3. This time-dependent degradation means that a patient with a 20-year history faces far more severe microstructural bone damage than a newly diagnosed individual, regardless of current HbA1c levels. Failing to account for disease duration may therefore bias risk estimates across different clinical stages. Second, the researchers did not differentiate between specific types of anti-diabetic medications (ADMs), thereby overlooking their heterogeneous effects on bone metabolism and fall risk. Clinical evidence suggests that while thiazolidinediones and specific SGLT2 inhibitors (e.g., canagliflozin) increase fracture risk4,5 metformin remains bone-neutral or even protective6. Furthermore, insulin and sulfonylureas are associated with increased fracture risk, likely mediated by hypoglycemia-induced falls6,7. Falls are often the direct trigger of hip fractures in the elderly, carrying a risk magnitude that far outweighs the impact of hyperglycemia-induced alterations in bone quality. Therefore, the heightened fracture risk reported in the diabetes group might be partially driven by medication-induced bone impairment and falls induced by drug-related hypoglycemia, rather than hyperglycemia alone. Failing to adjust for specific ADM regimens may mask these critical pharmacological influences and bias the causal interpretation. Finally, categorizing “hip fracture” as a single entity obscures potential differences in fracture subtypes and injury mechanisms. Current evidence suggests that intertrochanteric and femoral neck fractures may possess distinct pathophysiological profiles, with some studies indicating a higher prevalence of intertrochanteric fractures among the elderly and those in poorer physical condition8,9. We encourage the authors to consider these subtypes in future research to better elucidate their respective pathophysiological features. Additionally, without differentiating low-energy fragility fractures from high-energy trauma, the relative influence of physiological frailty versus external traumatic events remains confounded. In conclusion, while the study by Li et al provides compelling evidence for the joint impact of frailty and glycemic dysregulation on hip fracture risk, we suggest that future research should integrate data on diabetes duration, specific medication exposure, and fracture subtypes. Such a refined approach is essential to translate these epidemiological associations into clinical guidelines for fracture prevention in the aging diabetic population. The preparation of this manuscript adheres strictly to the 2025 TITAN Guidelines, ensuring appropriate disclosure and responsible application of generative AI tools10.