A) Any fracture B) Non-vertebral fracture C) Vertebral fracture

Supplementary Fig 1. Differences in plasma Dickkopf-1 levels following adjustment for confounders according to osteoporotic fracture (OF) status. Estimated means with 95% confidence intervals were generated and compared using analysis of covariance (ANCOVA). The multivariable adjustment factors were age, years since menopause, body mass index, current smoking status, alcohol intake (≥3 U/day), regular outdoor exercise (≥30 min/day), and parental history of OF.

Supplementary Table 1. Multiple logistic regression analyses to determine whether plasma Dickkopf-1 levelsmay predict osteoporotic fracture

A) Any fracture

ORs (95% CIs) per SD decrement
in plasma Dickkopf-1 concentration / P value
Unadjusted / 1.05 (0.80–1.38) / 0.743
Multivariable adjustment / 1.01 (0.76–1.34) / 0.962
Additional adjustment for TH BMD / 0.99 (0.75–1.32) / 0.981

B) Non-vertebral fracture

ORs (95% CIs) per SD decrement
in plasma Dickkopf-1 concentration / P value
Unadjusted / 1.11 (0.79–1.57) / 0.545
Multivariable adjustment / 1.02 (0.71–1.47) / 0.913
Additional adjustment for TH BMD / 1.01 (0.70–1.45) / 0.960

C) Vertebral fracture

ORs (95% CIs) per SD decrement
in plasma Dickkopf-1 concentration / P value
Unadjusted / 0.98 (0.71–1.35) / 0.898
Multivariable adjustment / 0.94 (0.68–1.32) / 0.728
Additional adjustment for LS BMD / 1.16 (0.76–1.75) / 0.498

The multivariable adjustment factors are age, years since menopause, body mass index, current smoking status, alcohol intake (≥3 U/day), regular outdoor exercise (≥30 min/day), and parental history of osteoporotic fracture. OR, odds ratio;CI, confidence interval; SD, standard deviation; TH BMD, total hip bone mineral density; and LS BMD, lumbar spine bone mineral density.

Supplementary Table 2. Multiple logistic regression analyses to generate odds ratios of covariates included in the multivariable adjustment model for any prevalent osteoporotic fracture

Covariates / Odds ratios (95% CIs) / P value
Total hip BMD (SD increment) / 0.76 (0.55–1.05) / 0.094
Age / 1.09 (0.98–1.21) / 0.108
Year since menopause / 1.04 (0.95-1.13) / 0.432
Body mass index / 0.95 (0.86-1.05) / 0.348
Current smoking / 0.71 (0.20-2.50) / 0.590
Alcohol intake / 1.14 (0.62-2.09) / 0.672
Exercise status / 0.59 (0.26-1.35) / 0.211
Parental history of OF / 3.16 (1.41–7.11) / 0.005

Values in bold indicate statistically significant values.CI, confidence interval; SD, standard deviation; BMD, bone mineral density; and OF, osteoporotic fracture.

Supplementary Table 3.Multiple linear regression analyses to determine the association of plasma Dickkopf-1 concentrations with BMD values at different skeletal sites and bone turnover markers

Dependent variables / Unadjusted / Multivariable adjustment
*β / P value / †R2 / β / P value / R2
Lumbar spineBMD / ˗0.013 / 0.862 / 0.001 / 0.001 / 0.992 / 0.104
Total hip BMD / 0.021 / 0.769 / 0.001 / 0.048 / 0.454 / 0.210
Femoral neckBMD / 0.027 / 0.704 / 0.001 / 0.057 / 0.384 / 0.170
TrochanterBMD / 0.017 / 0.810 / 0.001 / 0.040 / 0.540 / 0.188
CTX / -0.023 / 0.767 / 0.001 / -0.041 / 0.606 / 0.050
BSALP / -0.108 / 0.189 / 0.012 / -0.103 / 0.215 / 0.067

*Standardized coefficient. †R2 means the proportion of the response variable variation that is explained by a linear model. The enter method was applied to this model with the BMD value at each skeletal site serving as the dependent variable, and the plasmaDickkopf-1 concentration serving as the independent variable.The multivariable adjustment factors are age, years since menopause, body mass index, current smoking status, alcohol intake (≥3 U/day), regular outdoor exercise (≥30 min/day), and parental history of osteoporotic fracture. BMD, bone mineral density; CTX, C-terminal telopeptide of type I collagen; andBSALP, bone-specific alkaline phosphatase.

1