Bone density measurements were performed by dual energy X-ray absorptiometry (DXA). Lumbar spine scan included L1 through L4 vertebrae. All DXA scans were submitted to a central imaging vendor for analysis. A mixed effect model for repeated measurements (MMRM) was fitted to the changes from baseline in LS-BMD for all post-baseline time points up to Week 52. Values at Week 52 were estimated from the model and are presented in the table.

Bone density measurements were performed by DXA. Lumbar spine scan included L1 through L4 vertebrae. All DXA scans were submitted to a central imaging vendor for analysis. A MMRM was fitted to the changes from baseline in LS-BMD for all post-baseline time points up to Week 52. Missing values were assumed to be missing at random (MAR) using the MMRM model. Values at Week 52 were estimated from the model and are presented in the table.

Carboxy-terminal crosslinked telopeptides of type I collagen (CTX) is a bone resorption biomarker. Serum CTX concentration-time data were analyzed by non-compartmental methods. The AUEC of baseline corrected serum CTX concentrations (% change from baseline) was calculated using the linear trapezoidal method. Values below the lower limit of quantification (LLOQ) were imputed with the actual value for the LLOQ.

Serum denosumab concentration-time data were analyzed by non-compartmental methods. Missing denosumab serum concentrations or concentrations below the LLOQ were not imputed and handled as missing values, except for the pre-dose sample which were treated as zero.

Serum denosumab concentration-time data were analyzed by non-compartmental methods. Missing denosumab serum concentrations or concentrations below the LLOQ were not imputed and handled as missing values, except for the pre-dose sample which were treated as zero. The linear-up log-down trapezoidal method was used for the AUCinf calculation.