Association of high-density lipoprotein cholesterol with site- specific bone mineral density T-scores in patients with type 2 diabetes aged 50 years or older: a retrospective cross-sectional study

Background The association between lipid metabolism and bone mineral density remains controversial, particularly in patients with type 2 diabetes mellitus (T2DM). This study aimed to investigate the relationship between HDL-C and site-specific bone mineral density T-scores in patients with T2DM aged 50 years or older. Methods This retrospective cross-sectional study included 2,568 hospitalized patients with T2DM who underwent bone mineral density assessment between February 2022 and December 2025. Patients were classified into normal BMD, osteopenia, and osteoporosis groups according to T-score criteria. Correlation analyses were performed to evaluate associations between lipid parameters and site-specific T-scores. Multivariable linear regression models were used to assess the independent association between HDL-C and lumbar spine T-score. Restricted cubic spline analysis and subgroup analyses were further performed. Results Among the 2,568 patients, 1,138 had normal BMD, 970 had osteopenia, and 460 had osteoporosis. HDL-C and ApoA1 levels increased progressively across worsening BMD categories. Correlation analysis showed that HDL-C was negatively correlated with T-scores at all skeletal sites, with the strongest correlation observed for lumbar spine T-score. In multivariable linear regression analysis, higher HDL-C was independently associated with lower lumbar spine T-score. This association remained significant after full adjustment for demographic characteristics, lifestyle factors, metabolic and clinical variables, lipid-lowering therapy, and bone-related medication use (β = -0.79, 95% CI: -1.22 to -0.35, P < 0.001). Site-specific analyses also showed inverse associations between HDL-C and total hip, femoral neck, and Ward’s triangle T-scores. Restricted cubic spline analysis suggested an approximately linear inverse association between HDL-C and lumbar spine T-score, without evidence of nonlinearity. Subgroup analyses showed that the inverse association was more pronounced among patients with BMI < 24 kg/m², with a significant interaction by BMI. Conclusion Higher HDL-C levels were independently associated with lower lumbar spine T-score in patients with T2DM aged 50 years or older. This association was approximately linear and appeared stronger among patients with lower BMI. HDL-C may serve as a lipid-related marker associated with reduced lumbar spine bone density in older patients with T2DM, although causal inference cannot be established due to the cross-sectional study design.