The association of genetically predicted type 2 diabetes with fracture and BMD by using different MR methods. (A) the genetically predicted type 2 diabetes and fracture. (B) the genetically predicted type 2 diabetes and BMD. (C) visualized the association of lead SNPs for type 2 diabetes with the risk of fracture. Abbreviations: wGRS, weighted genetic risk score; MR-PRESSO, MR pleiotropy residual sum and outlier.

The venn diagrams, conditional Q-Q plots and genetic variants jointly associated with type 2 diabetes and fracture/BMD at conjunctional false discovery rate (conjFDR) < 0.05. (A) the shared number of variants between type 2 diabetes and fracture. (B) the conditional Q-Q plot of fracture given association with type 2 diabetes at the level of p ≤ 0.1, p ≤ 0.01, p ≤ 0.001. (C) the shared genetic loci between type 2 diabetes and fracture. (D) the shared number of variants between type 2 diabetes and BMD. (E) the conditional Q-Q plot of BMD given association with type 2 diabetes at the level of p ≤ 0.1, p ≤ 0.01, p ≤ 0.001. (F) the shared genetic loci between type 2 diabetes and BMD.

The distinct signal (RSPO3) shared by type 2 diabetes and fracture. (A) the regional plot of the association of type 2 diabetes and RSPO3 gene expression (adipose subcutaneous) within hg19: chr6:127189749-127689749 (RSPO3 gene region ± 250kb window). (B) the regional plot of the association of fracture and RSPO3 gene expression (adipose subcutaneous) in the same region. (C) the association of genetically predicted RSPO3 gene expression with type 2 diabetes and fracture risk. (D) the association of genetically predicted RSPO3 gene expression with BMI, waist circumference, waist-hip ratio and MRI-derived visceral adipose. (F) the association of genetically predicted circulating RSPO3 with type 2 diabetes and fracture risk.

The regression between type 2 diabetes and fracture/BMD in observational study. (A) the relationship between type 2 diabetes and fracture in different sites including all fracture, weight-bearing bones and other bones. Model 0 adjusted for reference age, sex, BMI, physical activity, fall history, HbA1c and medication treatments; Model 1 adjusted for a+ BMD. (B) the stratified analyses between type 2 diabetes, and fracture based on the five risk factors secondary to the diseases adjusted for the age and sex. (C) the relationship between type 2 diabetes and BMD and the stratified analyses based on the five risk factors secondary to the diseases adjusted for the age and sex. The 5 risk factors were: (1) BMI≤25kg/m2; (2) no physical activity; (3) falls in the last year; (4) HbA1c≥47.5mmol/mol; (5) antidiabetic medication treatments.

Flow chart of the overall study design.

The regression between type 2 diabetes GRS and fracture risk/BMD in male and female. Models were adjusted for reference age, sex, BMI, physical activity, fall history, HbA1c and medication treatments.

The regression between type 2 diabetes and fracture risk/BMD in male and female in observational study. Model 0 adjusted for reference age, sex, BMI, physical activity, fall history, HbA1c and medication treatments; Model 1 adjusted for a+ BMD.

The stratified analyses between type 2 diabetes and fracture/BMD based on the five risk factors (BMI≤25kg/m2, no physical activity, falls in the last year, HbA1c≥47.5mmol/mol and antidiabetic medication treatments) in male and female. Models were adjusted for the age and sex.

The matching information of 14,860 patients with fracture.