Assessment of fracture risk by FRAX model in older adults with type 2 diabetes: a cross-sectional study in China

Background: The FRAX model is an effective tool to assess fracture risk, but its application has not been assessed in patients with type 2 diabetes in Chinese mainland. We investigated FRAX-estimated fracture risk in older patients with type 2 diabetes mellitus (T2D) compared with control subjects. Methods: In our study, we assessed the FRAX scores of 267 T2D and 359 non-diabetic subjects from Tongji Hospital, Tongji University School of Medicine. We tested bone mineral density (BMD) and calculated FRAX scores. Binary regression analysis was used to evaluate the risk factors for high risk fracture prediction by FRAX model. Results: The BMI (Body Mass index), WHR (Waist-hip ratio), frequency of smoking, and alcohol consumption were significantly higher in T2D. T2D had significantly elevated BMC (Bone mineral content), T scores, Z scores in FN (Femoral neck) and in LS (lumber spine) (T score: -1.4 +/- 1.6 vs. -2.1 +/- 1.4, -1.7 +/- 1.1 vs. -2.0 +/- 1.0, P<0.001; Z score: -0.1 +/- 1.5 vs. -0.6 +/- 1.3; -0.4 +/- 1.0 vs. -0.7 +/- 1.0, P<0.001). T2D had lower FRAX-estimated probability of both major osteoporotic fracture (MOF) and hip fracture (HF) by FRAX-BMI model (4.27 +/- 2.84% vs. 5.14 +/- 2.92%, P<0.001, and 1.42 +/- 1.54% vs. 1.83 +/- 2.23%, P<0.001, respectively). T2D had lower FRAX-estimated probability of MOF by FRAX-BMD model (5.28 +/- 4.41% vs. 6.30 +/- 5.10%, P<0.001). When grouping by BMI, T2D had lower FRAX scores of MOF by FRAX-BMI. Binary regression analysis showed FN T score (B4.58, P<0.001), smoking (B1.489, P<0.001), family history of hip fracture (B1.993, P<0.001) and corticosteroids use (B2.886, P<0.001) contributed to high risk of HF. FN T score (B-5.313, P<0.001), smoking (B3.753, P<0.01), family history of hip fracture (B2.521, P<0.001) and previous history of fracture (B3.239, P<0.05) contributed to high risk of MOF. Presence of T2D was not a contributing factor to high risk of fracture. Conclusions: Older Chinese patients with T2D had lower mean FRAX scores than non-diabetic subjects. T2D was not a risk contributor to high risk fracture prediction by FRAX. FRAX tool underestimated fracture risk in T2D population. Patients with T2D may be considered treatment when FRAX score was below FRAX-based intervention threshold.