Association of different lipid measures with incident bone fractures: Tehran lipid and glucose study

To investigate the association between different lipid measures and long-term hospitalization-required incident fracture among Iranian men and women. A total of 3309 individuals aged ≥50 years (men = 1598) were included in the study. Multivariate Cox proportional hazard analyses were performed to assess the risk of incident fracture across quintiles, considering first quintile as reference, as well as for 1-standard deviation (SD) increase in each lipid measure, i.e. total cholesterol (TC), triglycerides (TG), low- and high-density lipoprotein cholesterol (LDL-C and HDL-C, respectively), non-HDL-C, and related indices (TG/HDL-C and TC/HDL-C). Covariates included age, body mass index, current smoking, type 2 diabetes mellitus, hypertension, lipid lowering-drugs, and steroid medications (for women). During a median follow-up of 18 years, incident fracture was observed in 201 cases (men = 87). In both gender, no linear association was found between different lipid measures and incident fracture. Among men, only the fourth quartile of TG was associated with lower risk of fracture in the age-adjusted analysis with the hazard ratio (HR) and 95% confidence interval (CI) of [0.45 (0.21-0.95)]. Among women, the age-adjusted HRs and 95% CIs for the second, third, fourth, and fifth quintiles of non-HDL-C were [0.46 (0.25-0.87)], [0.73 (0.42-1.25)], [0.90 (0.54-1.51)], and [0.52 (0.29-0.95)], respectively; the corresponding values in the multivariate model were [0.48 (0.26-0.90)], [0.76 (0.4-1.32)], [0.94 (0.56-1.58)], and [0.52 (0.28-0.95)], respectively. The second quintile of LDL-C was also associated with lower risk for incident fracture in the multivariate analysis [0.53 (0.29-0.98)]. Among Iranian women, a nonlinear association between non-HDL-C and LDL-C and incident fracture was found as the second and fifth quintile of the former and the second quintile of the latter were associated with about 50% lower risk of fracture. Generally, our findings did not support harmful impact of these lipid measures on incident fracture.