Skip to main content
. 2021 Nov 1;22:921. doi: 10.1186/s12891-021-04795-4

Table 2.

Binary logistic regression models for the contribution of musculoskeletal factors to frailty for men. Data presented as odds ratio (OR, 95% confidence interval (CI))

Exposures OR (95% CI) AUROC P-value
Femoral neck BMD
 Unadjusted 0.99 (0.70–1.40) 0.52 0.96
Ultrasound SOS
 Unadjusted 0.90 (0.74–1.10) 0.58 0.29
Ultrasound BUA
 Unadjusted 0.82 (0.62–1.08) 0.55 0.16
Ultrasound SI
 Unadjusted 0.73 (0.54–1.00) 0.57 0.05
 Adjusted- age, BMI 0.73 (0.53–1.01) 0.71 0.06
 Final Model * * *
ALMI
 Unadjusted 0.60 (0.44–0.81) 0.65 < 0.01
 Adjusted- age, fat mass index 0.50 (0.36–0.70) 0.75 < 0.01
 Final Model 1- age, fat mass index, SES 0.50 (0.35–0.70) 0.77 < 0.01
 Final model 2-age, fracture, SES, fat mass index 0.48 (0.34–0.68) 0.77 < 0.01
HGS
 Unadjusted 0.12 (0.07–0.23) 0.83 < 0.01
 Adjusted- age, height 0.12 (0.06–0.23) 0.84 < 0.01
 Final Model- age, height, SES 0.11 (0.06–0.22) 0.85 < 0.01

*Final model same as adjusted model

AUROC area under the receiver operational characteristic curve, BMD bone mineral density, BUA broad-band ultrasound attenuation, SOS speed of sound, SI stiffness index, ALMI appendicular lean mass index, HGS handgrip strength