Jiao Zhong_supplemental material.pdf

Diabetes mellitus (DM), osteoporosis (OP), and obesity (OB) are all complex diseases. OB is associated with both DM and OP, but it’s unclear whether OB mediates association between DM and OP.The study aimed to investigate the potential mediation effects of OB on association between DM and BMD by the causal inference tests (CIT). A total of 5,682 Chinese aged over 65 years were enrolled in an ongoing Osteoporosis Preventive Project. Obesity-related indexes, including body mass index (BMI), waist circumference, and waist circumference-hip circumference-ratio (WHR), and bone mineral density (BMD) at total hip (TH) and femur neck (FN) were measured. OP was diagnosed with a T-score of BMD less than or equal to −2.5, and osteopenia was diagnosed with a T-score of BMD between −2.5 and −1.0. Subjects with a T-score of BMD −1.0 or higher were considered as normal according to the criteria established by WHO 【1】. OB was diagnosed and subjects were categorized as follows: normal-weight (Group1: BMI≤23.9 kg/m2), overweight (Group2: 24.0 kg/m2≤BMI≤27.9 kg/m2), and obese (Group3: BMI≥28.0 kg/m2). CIT among the three diseases/traits of interest requires the following four criteria to be met simultaneously:(1) The causal factor A is associated with the outcome C, (2) the causal factor A is associated with the mediator B with adjustment for outcome C, (3) the mediator B is associated with the outcome C after adjusting for the causal factor A, and (4) the causal factor A is independent of the outcome C after adjusting for the mediator B 【2】. CIT was used to identify potential mediators underlying the association between DM and BMD/OP.<br>