Background: Development of assessment tool for fracture risk is an urgent task, because bone mineral density (BMD) is less useful for evaluating fracture risk in type 2 diabetes mellitus (T2DM).
Subjects and methods: In total, 808 T2DM patients were enrolled in this cross-sectional study. To develop a scoring assessment tool using clinical risks for vertebral fracture (VF), we evaluated which variables were associated with VF by logistic regression analysis, and categorized these variables based on cut-off values obtained by using receiver operating characteristic (ROC) curves. For calculation of the score, the relative weight of the factors was determined, and a tentative score was assigned. Then, cut-off point of the score was examined to predict VF.
Results: Logistic regression analyses showed that age, diabetes duration, body mass index (BMI), serum albumin, and T score at femoral neck (FN-T score) were associated with VF risk. Parameter estimates for each risk factor obtained by logistic analyses were converted to risk scores (maximum score 23). ROC analysis showed that 8.5 was the cut-off value for detecting VF. Multiple logistic regression analysis adjusted for confounding factors showed that score ≥9 was significantly associated with an increased risk of prevalent VF (odds ratio 1.99, 95% confidence interval 1.22-3.24, p = 0.006).
Conclusions: This is the first study to show that a scoring assessment tool using age, duration of diabetes, BMI, serum albumin, and FN-T score is useful to estimate VF risk in patients with T2DM, being more sensitive than BMD alone in detecting bone fragility.
Keywords: Assessment tool; Bone mineral density; Type 2 diabetes mellitus; Vertebral fracture.
Copyright © 2019 Elsevier Inc. All rights reserved.