Prediction model for low bone mass mineral density in type 2 diabetes: an observational cross-sectional study.

Abstract

Purpose: Considering the prevalence of type 2 diabetes (T2D), osteoporosis should be considered a serious complication. However, an effective tool for the assessment of low bone mass mineral density (BMD) in T2D patients is not currently available. Therefore, the aim of our study was to establish a simple-to-use risk assessment tool by exploring risk factors for low BMD in T2D patients.

Methods: This study included 436 patients with a low BMD and 381 patients with a normal BMD. Multiple logistic regression analysis was performed to evaluate risk factors for low BMD in T2D patients. A nomogram was then developed from these results. A receiver operating characteristic (ROC) curve, calibration plot, and goodness-of-fit test were used to validate the nomogram. The clinical utility of the nomogram was also assessed.

Results: Multivariate logistic regression indicated that age, sex, education, body mass index (BMI), fasting C-peptide, high-density cholesterol (HDL), alkaline phosphatase (ALP), estimated glomerular filtration rate (eGFR), and type I collagen carboxy terminal peptide (S-CTX) were independent predictors for low BMD in T2D patients. The nomogram was developed from these variables using both the unadjusted area under the curve (AUC) and the bootstrap-corrected AUC (0.828). Calibration plots and the goodness-of-fit test demonstrated that the nomogram was well calibrated.

Conclusions: The nomogram-illustrated model can be used by clinicians to easily predict the risk of low BMD in T2D patients. Our study also revealed that common factors are independent predictors of low BMD risk. Our results provide a new strategy for the prediction, investigation, and facilitation of low BMD in T2D patients.