The investigation has been conducted following the ethical standards of the Declaration of Helsinki and national and international guidelines. This study has been approved by the Ethics Committee of Xiangya Second Hospital, South China University, Changsha, China. All participants provided written informed consent.
This is a case-control study. Postmenopausal women (n = 434) were recruited from community medical centers in Changsha City, Hunan Province, China, in September 2017. These included 217 T2DM patients and 217 consecutive postmenopausal women without diabetes as a control group. All participants had undergone baseline dual x-ray absorptiometry (DXA) of the lumbar spine, femoral neck and hip. The control and T2DM groups were matched for age and BMI. Menopause was deﬁned as the absence of menstrual cycles for over one year. Exclusion criteria included the use of antiosteoporosis medication, presence of a malignant tumor, acute infection, and severe renal dysfunction. All participants completed standard medical assessment questionnaires to collect data about their medical, social, and family history, and clinical risk factors of FRAX. These factors included a parental history of hip fracture, history of fragility fractures, exposure to systemic glucocorticoids for more than three months, smoking, high alcohol intake (an average of three or more units daily), presence of RA, secondary osteoporosis factors, adult osteogenesis deficiency, long-term untreated hyperthyroidism, hypogonadism or premature menopause (<45 years), chronic malnutrition or malabsorption, and chronic liver disease.
Height was measured to the nearest 0.5 cm, using a stadiometer mounted on a wall and pressing a horizontal plate onto the head to ﬂatten the hair. An electronic scale measured weight to the nearest 0.1 kg. The BMI was calculated by dividing the body weight by the squared height (kg/m2). The participants age was recorded.
DXA scans were performed for the left femoral neck, total hip, and lumbar spine following the manufacturer’s guidelines (Discovery Wi S/N87556, Hologic, USA), and they were measured by an experienced doctor. Osteoporosis was defined as a T-score ≤ −2.5 following the World Health Organization recommendations used to establish the BMD reference database by our group . Positive and lateral vertebral radiographs (T4 to L4) were acquired (Uni-Vision 61Y040, Shimadzu, Japan) to assess vertebral fractures, using a visual semi-quantitative method. Fractures were defined as a ratio of the anterior to posterior or middle to posterior vertebral height < 0.8, or the posterior vertebral height compared to that of an adjacent vertebra .
Base on the guidelines for diagnosis and treatment of primary osteoporosis in China, osteoporosis was diagnosed if a patient met any of the following criteria: 1) fragility fractures of the hip or vertebra; 2) T-score of less than −2.5 for BMD measured by DXA in the medial axial skeleton; 3) low bone mass based on the BMD results (−2.5 < T-score < −1.0) with fragile fractures of the proximal humerus, pelvis, or distal forearm . Diabetes diagnosis followed the guidelines for the prevention and control of T2DM in China (2017 Edition) .
The 10-year risk for MOF, including OHF, were calculated using the China FRAX® model, with the femoral neck T-score included in the calculations. We compared three alternative options to enhance the performance of FRAX in patients with diabetes. First, FRAX with RA input as a proxy for the effect of diabetes; this is justified by the similar weights assigned to RA and T2DM in the QFracture algorithm . Second, FRAX with the femoral neck T-score input lowered by 0.5 SD in T2DM patients; other studies have observed that women with and without T2DM and equivalent OHF risk differed in their T-score by approximately 0.5 SD . Third, FRAX with a 10-year higher age input for T2DM patients. This age correction is equivalent to an expected femoral neck BMD loss of 0.5 SD over ten years.
Continuous variables are presented as mean ± SD. Normal distribution of continuous variables was assessed separately for diabetic and nondiabetic participants by the Kolmogorov–Smirnov test. Categorical variables were analyzed by the Chi-squared test. We constructed ROC curves to determine the AUC and the corresponding 95% confidence intervals to identify the reference variable. The AUC is an overall estimate of the risk score accuracy in identifying the optimal choice . The AUCs were compared using Stata 14.0 (Stata Corp, College Station, Tx). The cutoff value for statistical significance was set at p < 0.05. The best cutoff point in the AUCs was estimated using the Youden index, and the accuracy of the thresholds was assessed by analyzing their sensitivity, specificity, and likelihood ratios. All data management and part of the analysis were performed using SPSS Statistics for Windows, Version 24.0 (SPSS Inc., Chicago, IL, USA).