Table 2. Logistic Regression Analysis Resultsa.
Possible risk factor (cutoff value) | Univariate analysis | Multivariate analysis | ||
---|---|---|---|---|
OR (95% CI) | P value | OR (95% CI) | P value | |
Age (72), y | 0.44 (0.12-1.57) | .21 | NA | NA |
Height (153.0), cm | 1.17 (0.33-4.13) | .81 | NA | NA |
Body weight (56.6), kg | 6.00 (1.20-29.97) | .03 | 4.44 (1.04-18.96) | .04 |
BMI (18.5) | 1.63 (0.29-9.30) | .58 | NA | NA |
VCF | ||||
Prevalent | 2.15 (0.25-18.82) | .49 | NA | NA |
Prior multiple | 2.23 (0.60-8.23) | .23 | NA | NA |
Denosumab treatment duration, y | ||||
≥2.5 vs <2.5 | 3.23 (0.89-11.75) | .08 | NA | NA |
≥3.0 vs <3.0 | 8.17 (1.99-33.58) | .004 | 6.91 (1.75-27.28) | .006 |
Bone turnover marker serum level, ng/mL | ||||
CTX (0.30) | 0.69 (0.20-2.44) | .56 | NA | NA |
P1NP (23.97) | 0.37 (0.09-1.52) | .17 | NA | NA |
Osteoporosis before trial enrollment | 0.45 (0.12-1.66) | .23 | NA | NA |
Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); CTX, C-terminal telopeptide; NA, not applicable; OR, odds ratio; P1NP, N-terminal propeptide of type 1 collagen; VCF, vertebral compression fracture.
Results are presented for the univariate analysis and the multivariable logistic regression using stepwise selection, which was conducted only for risk factors with P < .10 in the univariate analysis. With regard to denosumab treatment duration, there were 2 competing factors; only the more statistically significant factor was incorporated by the model.